Ecosystem changes through the Permian–Triassic and Triassic–Jurassic critical intervals: Evidence from sedimentology, palaeontology and geochemistry

The Permian–Triassic and Triassic–Jurassic critical intervals are among the most significant ecological upheavals in the Phanerozoic. Both evolutionary junctures are characterized by environmental deterioration associated with a marked biodiversity decline. In this study, Permian–Triassic and Triassic–Jurassic boundary sections from South China and the Northern Calcareous Alps were investigated. In order to reconstruct the interplay between biotic and abiotic processes, a multifaceted approach that included optical microscopy, X‐ray diffraction, Raman spectroscopy, stable carbon isotopes and lipid biomarkers was employed. The lower parts of these two sections are similar as both consist of limestone with abundant fossils of eukaryotic organisms. However, the Permian–Triassic record is dominated by dasyclad green algae and fusulinid foraminifera, while the Triassic–Jurassic record is typified by corals and coralline sponges. Moving upward, both sections consist mainly of micrite and marl. Concerning the Permian–Triassic section, it transits to volcanic ash intercalated by a distinct limestone bed with abundant calcispheres (tentatively attributed to ancestors of dinoflagellates). The Triassic–Jurassic section does not provide direct evidence for volcanic activity, but also becomes rich in calcisphere‐type cysts towards the top. Additionally, the section preserves abundant 4‐methyl sterenes (diagnostic for dinoflagellates) and C37–39 n‐alkanes (indicative for haptophytes). Hence, both critical intervals were associated with marked blooms of (ancestral) dinoflagellates and haptophytes (for example, coccolithophorids). These blooms were followed by ecological lag‐phases, as indicated by low carbonate contents and scarce fossils which only increased further up the sections. For both critical intervals, it is commonly assumed that the formation of voluminous volcanic provinces (Siberian Traps and Central Atlantic Magmatic Province, respectively), as well as associated processes (for example, burning of organic‐rich sediments such as coal), resulted in ecological devastation. However, results suggest that volcanism also had a positive effect on certain planktonic primary producers such as dinoflagellates and haptophytes, perhaps by delivering essential nutrients.


INTRODUCTION
Most Phanerozoic extinction events are identified by quantitative analyses using databases of marine fossil organisms (e.g. Raup, 1979;Sepkoski et al., 1981;Raup & Sepkoski, 1982;Bambach, 2006;Alroy et al., 2008;Fan et al., 2020). Two of the most severe Phanerozoic mass extinctions coincide with ecological upheavals during the Permian-Triassic (P-T) and Triassic-Jurassic (T-J) critical intervals. Both events have been intimately associated with the formation of large continental flood basalt provinces: the Siberian Traps and the Central Atlantic Magmatic Province (CAMP), respectively (Whiteside et al., 2010;Blackburn et al., 2013;Burgess & Bowring, 2015;van de Schootbrugge & Wignall, 2016;Burgess et al., 2017;Davies et al., 2017;Broadley et al., 2018). In addition to the relative timing of the events, the potential relationship between volcanism and ecosystem change is evidenced by widespread occurrences of volcanic ash layers and mercury anomalies in marine sedimentary rocks (Thibodeau et al., 2016;Percival et al., 2017;Wang et al., 2018).
Geochemical evidence encoded in sedimentary rocks contributes to the understanding of ecological upheavals during the P-T and T-J intervals. For instance, significant shifts in the stable carbon isotopic composition of P-T carbonates are proposed to reflect significant perturbances in the global carbon cycle at that time (Payne et al., 2004). Similarly, profound shifts in the stable sulphur isotopic composition of sulphate and carbonate-associated sulphate might indicate that the global biogeochemical sulphur cycle was affected as well (Luo et al., 2010;Song et al., 2014;Bernasconi et al., 2017). During the T-J crisis, major biogeochemical cycles were also strongly perturbed as reflected by significant carbon and sulphur isotopic fluctuations (Williford et al., 2009;Ruhl et al., 2010. During the P-T and T-J crises, many parts of the oceans perhaps became oxygen-poor or even anoxic, also in the photic zone (Brennecka et al., 2011;Richoz et al., 2012;Jaraula et al., 2013;Blumenberg et al., 2016;Huang et al., 2017;Zhang et al., 2018), and seawater pH decreased (Payne et al., 2010;Greene et al., 2012;Hinojosa et al., 2012;Clarkson et al., 2015). However, one major difference between the two critical intervals might have been seawater temperatures: P-T oceans were suggested to be lethally hot (up to 36°C) based on the d 18 O apatite conodont record Sun et al., 2012), perhaps driving the expansion of anoxic conditions (Penn et al., 2018), whilst T-J oceans are characterized by mild to warm temperatures (ca 13 to 30°C) as indicated by clumped-isotope palaeothermometry studies on microbialites (Petryshyn et al., 2020). During the P-T crisis, hypercapnia (i.e. abnormally high CO 2 levels in the blood of organisms, causing physiological problems) might have also driven ecosystem change (Knoll et al., 2007), which is supported by high atmospheric carbon dioxide contents (Wu et al., 2021).
This study tracks ecosystem changes through the P-T and T-J critical intervals in South China and the Northern Calcareous Alps, respectively. Following a geobiological approach, the study combined techniques from sedimentology, palaeontology and biogeochemistry, including optical microscopy, X-ray diffraction, Raman spectroscopy, stable carbon isotopes and lipid biomarkers. The detailed comparison of records from both localities allows identification of fundamental similarities and differences between the P-T and T-J events, which is of great significance to the ongoing discussion of possible causes and triggers of mass extinctions.

GEOLOGICAL SETTINGS South China
South China is a key area for investigating the P-T critical interval because of the abundance of easily accessible outcrops and sections comprising well-preserved strata. This study focuses on the Xiakou section (N 31°06.874 0 , E 110°48.209 0 ), close to Xiakou town, Hubei Province (Fig. 1A).
In this area, a marine transgression at the end of the Permian resulted in a transition from shallow open platform (Changhsingian, Late Permian) to deeper shelf basin environments (Induan, Early Triassic) ( Fig. 1A; Feng et al., 1997;Luo et al., 2009). Stratigraphically, the Xiakou section comprises the Changxing and Dalong formations (Changhsingian, Upper Permian) as well as the Daye Formation (terminal Changhsingian, Upper Permian to Induan, Lower Triassic). The lowest occurrence of the conodont Hindeodus parvus at the base of bed T 1 in the Daye Formation marks the beginning of the Triassic ( Fig. 2; Zhao et al., 2013). The base of bed T 1 can thus be correlated with the base of bed 27c of Meishan Section D (Changxing County, Zhejiang Province, China), the Global Stratotype Section and Point (GSSP) of the P-T boundary (Yin et al., 2001;Zhang et al., 2009). The Meishan record comprises palaeontological evidence for two mass extinction episodes (ME 1 and ME 2) across the P-T boundary (Song et al., 2013), which are correspondingly identified in the Xiakou section ( Fig. 2; Pei et al., 2021).

Northern calcareous Alps
During the T-J critical interval, the area of today's Northern Calcareous Alps was situated at the margin of the Meliata Ocean, which was part of the Neotethys (Stampfli & Borel, 2002). It is a key region for studying the T-J boundary because sedimentary successions represent a variety of different palaeoenvironments (e.g. Fabricius, 1966;Kuss, 1983;Mandl, 2000;McRoberts et al., 2012). The Lahnewies Syncline, located in the Ammer Mountains near Garmisch-Partenkirchen, Bavaria, belongs to the northern part of the upper Austroalpine Lechtal nappe and is traversed by various longitudinal and transverse faults (Kuhnert, 1967;Karl et al., 2014;Hornung & Haas, 2017). The stratigraphic section investigated herein is composed of two parts that are both situated in the western limb of the Lahnewies Syncline. More precisely, the main part is exposed close to the Lahnewies Creek (N 47°31.578 0 , E 11°04.002 0 ) (Fig. 1B), while the second part is exposed in the Nudelgraben Creek (N 47°29.045 0 , E 10°56.542 0 ). The T-J Lahnewies Syncline section shares many  Pei et al., 2021), between the town of Xiakou and the village of Jianyangping (Hubei province, South China). The strata were indicated after Luo et al. (2009). (B) Location of the Triassic-Jurassic (T-J) Lahnewies Syncline section, the main part exposed close to the Lahnewies creek near Garmisch-Partenkirchen, Southern Germany. J, Jurassic; T, Triassic; P, Permian; D, Devonian; S, Silurian. similarities with the GSSP section at Kuhjoch in the Karwendel-Syncline (Hillebrandt et al., 2007(Hillebrandt et al., , 2013. Detailed descriptions on the stratigraphy of the Upper Triassic in the Northern Calcareous Alps have been provided elsewhere (Fig. 3;Mandl, 2000;Krystyn et al., 2005;McRoberts et al., 2012;Karl et al., 2014). Briefly, sections across the T-J boundary typically begin with dolomites of the Hauptdolomit Formation (middle to upper Norian, Upper Triassic), which represented inter-supratidal environments on the Dachstein carbonate platform. The Norian-Rhaetian Dachstein Formation also includes 'Oberrhaetkalk' of the K€ ossen Basin and Sevatian Plattenkalk unit as a shallow subtidal facies between supratidal to subtidal back-reef facies of the Dachstein Reef complex. The Hauptdolomit Formation is locally overlain by limestones of the Plattenkalk unit (Sevatian/upper Norian, Upper Triassic), which formed in slightly deeper, shallow subtidal environments on the platform. Both facies develop into the K€ ossen Formation (Rhaetian, Upper Triassic), which represents an open marine lagoonal back reef environmentthe bathymetrically deepest part of the Dachstein carbonate platform. To the south, the Dachstein carbonate platform was bordered by deep water environments of the Hallstatt Basin, which was part of the Meliata Ocean. These environments are represented by fore reef marls of the Zlambach Formation, which is equivalent to the lagoonal K€ ossen Formation ( Fig. 3; Fabricius, 1966).
The K€ ossen Formation can be subdivided into the Hochalm and Eiberg members. The Eiberg Member is terminated by ca 50 to 100 cm thick T-Bed, which is strongly enriched in organic matter at its top. The T-Bed is of crucial stratigraphic significance because it can be traced throughout the area and contains terminal Triassic index fossils such as the ammonite Choristoceras marshi and the conodont Misikella posthersteini (e.g. Kozur & Mock, 1991;Krystyn, 2008).

Fieldwork and petrography
Field work in South China and the Northern Calcareous Alps has been conducted over several years and included careful observation of geological, sedimentological and palaeontological features as well as sampling. This was particularly important in the case of the Lahnewies Syncline, as sections and stratigraphic units are commonly tectonically separated. Details about the Xiakou section can be found in Pei et al. (2021).
Thin sections and smear slides were prepared for the analysis of microfacies and nannoplankton-type fossils (for more details on the latter see Janofske, 1987). Microfacies analysis was performed with a Zeiss SteREO Discovery.V12 stereomicroscope and a Zeiss AXIO Imager. Z1 microscope (Carl Zeiss AG, Jena, Germany). For fluorescence microscopy, the Zeiss AXIO Imager. Z1 microscope was equipped with a 10 AF488 filter (excitation wavelength = BP 450 to 490 nm, emission wavelength = BP 515 to 565 nm). Cathodoluminescence microscopy was conducted with a Zeiss Axiolab microscope combined with a Citl CCL 8200 Mk3A cold-cathode system (operating voltage of ca 15 kV; electric current of ca 250 to 300 lA) (Cambridge Image Technology Limited, Hatfield, UK). For field emission scanning electron microscopy (FE-SEM), a Carl Zeiss LEO 1530 Gemini system was used.
Raman spectroscopy was performed with a WITec alpha300 R fibre-coupled ultra-high throughput spectrometer (532 nm excitation laser, operated at 20 mW, 1009 long working distance objective with numerical aperture of 0.75 and a 300 g mm À1 grating) (Oxford Instruments Plc, Abingdon, UK). Centred at 2220 cm À1 , the spectrometer covered a spectral range from 68 to 3914 cm À1 with a spectral resolution of 2.2 cm À1 . Each spectrum was collected by two accumulations, with an acquisition time of 2 s. The Raman spectra were analysed via the WITec Project software. Automated cosmic ray correction, background subtraction and fitting using a Lorentz function were performed.
Total organic carbon (TOC) and total inorganic carbon (TIC) contents were measured with a Leco RC612, a Leco CS230 (Leco Corporation, St. Joseph, MI, USA) and an Euro EA 3000 elemental analyser (HEKAtech GmbH, Wegberg, Germany). All values are reported in weight % (wt.%) of bulk sedimentary rock. TOC and TIC contents of P-T samples were partially published in Pei et al. (2021).
Programmed pyrolysis of the P-T samples was carried out at Applied Petroleum Technology (Oslo, Norway), using a HAWK instrument (Wildcat Technologies, Humble, Texas, USA). The Late Triassic samples were run at the Federal Institute for Geosciences and Natural Resources (BGR) in Hannover, Germany, on a Rock-Eval-6 analyser (Vinci Technologies, Nanterre, France). Following established protocols (Espitali e et al., 1977;Lafargue et al., 1998), the temperature was ramped from 300°C (held for 3 min) to 650°C at 25°C min À1 . Initial sample weights were between 10 mg and 200 mg, depending on the expected hydrocarbon yields from labile kerogen (S2, see below) to prevent oversaturation of the flame ionization detector (FID). The analyses included quantities of free hydrocarbons (S1, mg HC g À1 rock), S2 (mg HC g À1 rock), temperatures at maximum yields of S2 hydrocarbons (T max ,°C), and the CO 2 generated from organic carbon at higher temperatures up to 650°C (S3, mg CO 2 g À1 rock). Based on these data and corresponding TOC data, the hydrogen index (HI, S2/TOC 9 100), the oxygen index (OI, S3/TOC 9 100) and the production index [PI, S1/(S1 + S2)] were calculated. For P-T samples, the data were already published in Pei (2022).

Lipid biomarkers
Lipid biomarkers were extracted and analysed at BGR. Only Late Triassic samples are considered in this study, as organic matter in P-T samples from Xiakou has a too high thermal maturity (for more details see Discussion) (Pei, 2022).
Details on the extraction, fractionation and subsequent analysis with gas chromatography (GC) -FID or GC-mass spectrometry (MS) can be found in Blumenberg et al. (2019). Briefly, organic compounds were extracted with dichloromethane (DCM) using an accelerated solvent extraction device (Dionex ASE-350, Thermo Fisher Scientific, Waltham, MA, USA). About 10 g of each sample was diluted with 22 g sea sand (Carl Roth GmbH + Co. KG, Karlsruhe, Germany) and extracted and annealed, 4 h at 400°C within an extraction cell. Organic extracts were collected automatically, and elemental sulphur was subsequently removed by granulated, activated copper (10% HCl at 60°C for approximately 1 h). The extracts were fractionated by mid-pressure liquid chromatography into an aliphatic, an aromatic and a polar fraction using isohexane, isohexane/dichloromethane (2/1; v/v) and dichloromethane/methanol (1/1; v/v), respectively. Only the aliphatic fraction was analysed for this study.
Stable carbon isotope analyses (d 13 C carb and d 13 C org ) The stable carbon isotopic composition of carbonates was analysed in the Isotope Geology Department at the Geoscience Center at the Georg-August-Universit€ at G€ ottingen (Germany). Briefly, ca 100 to 600 lg sample material was obtained with a high-precision drill to ensure sampling of individual carbonate microfacies and then measured with a Thermo Scientific Kiel IV carbonate device coupled to a Finnigan DeltaPlus gas isotope mass spectrometer (Thermo Fisher Scientific). Results are reported as delta values relative to Vienna Pee Dee Belemnite (VPDB) reference standard (d 13 C carb ). Reproducibility was repeatedly tested and generally better than 0.1&. Please note that d 13 C carb values of P-T samples were in part already published in Pei et al. (2021).
The stable carbon isotopic composition of organic matter was analysed in the Center for Stable Isotope Research and Analysis (KOSI) at the Georg-August-Universit€ at G€ ottingen (Germany). Briefly, ca 0.5 to 25 mg of decalcified sample material was measured with an Euro EA 3000 elemental analyser (HEKAtech GmbH, Wegberg, Germany) coupled to a Finnigan Delta V Advantage isotope mass spectrometer (Thermo Fisher Scientific). Results are reported as delta values relative to Vienna Pee Dee Belemnite (VPDB) reference standard (d 13 C org ). Acetanilide (d 13 C = À29.6 &; SD = 0.1) was used for internal calibration. Average standard deviation of d 13 C org values was 0.3&. More information on the protocol is provided in Werner et al. (1999), and Langel & Dyckmans (2014. It is important to note that d 13 C org values of P-T samples reported in this study are not identical to those reported in Pei et al. (2021) because they were re-measured for this study.

Field observations and microfacies analyses
The studied interval of the Xiakou section is about 9 m thick, and includes the Changxing, Dalong and Daye formations (Fig. 2). As detailed below, the section mainly comprises grey to black limestones, intercalated with black mudstones (Pei et al., 2021). Several volcanic ash beds/layers occur throughout the section (for example, beds/ layers P 15-1 , P 15-3 , T 2 , T 4 , T 13 and T 15 ) ( Fig. 2). As indicated by XRD analyses, the ash mainly consists of quartz and illite-smectite, agreeing with findings of previous studies (Table S1; Hong et al., 2011;Gao et al., 2013). Notably, some of the ash layers occur at the top of the Dalong Formation, close to the P-T boundary (layers P 15-1 and P 15-3 ). These greenish-yellowish layers are 11 cm and 13 cm thick, respectively. Between them, a 14 cm thick, black medium-bedded limestone (layer P 15-2 ) containing abundant calcispheres ( Fig. 2) was intercalated.
In a previous study, microfacies types (MF) were defined for the relevant part of the Xiakou section (Pei et al., 2021). Briefly summarizing, the uppermost part of the Changxing Formation consists of grey grainstones containing dasyclad green algae and fusulinid foraminifera (MF-1). The Dalong Formation mainly comprises black, laminated micrites with either dolomite rhombs (MF-2) or calcite pseudomorphs after gypsum and Rectocornuspira foraminifera (MF-3) (  MF-4a, black, laminated marls with abundant fossil debris, for example, thin-shelled bivalves (Bi) and small benthic foraminifera (Fo). (E) MF-4b, black, non-laminated marls with calcite and pyrite concretions (Co). Fragmented thin-shelled bivalves (Bi) and small benthic foraminifera (Fo) were noticed. (F) MF-7, grey floatstone to wackestone containing ammonoids (Am). Gastropods (Ga) and radiolarians (Ra) were displayed. well as black, non-laminated micrites with abundant calcispheres at the top (MF-5) ( Fig. 6A and B). The overlying Daye Formation consists of grey, non-laminated micrite with dolomite rhombs (MF-6) and grey, floatstones to wackestones containing ammonoids (MF-7) (Fig. 5F). The calcisphere-rich MF-5 is restricted to layer P 15-2 of the Dalong Formation and appears to be particularly interesting for several reasons. First, the matrix of MF-5 is characterized by copious peloidal textures ( Fig. 6A and B), which exhibits a relatively strong green fluorescence ( Fig. 7A and B) and a patchy cathodoluminescence ( Fig. 7C and D). Raman spectroscopy reveals that the matrix not only consists of calcite, but also comprises organic matter ( Fig. 8A and B). Second, it contains abundant calcispheres, which range from ca 50 lm to 300 lm in diameter and exhibit thin micritic walls. Three calcisphere types (types A to C) can be distinguished ( Fig. 6C to F). Calcispheres of type A and B enclose sparite with clear and indistinct grain boundaries, respectively ( Fig. 6C and D). The sparite within calcispheres of type B shows a relatively dull green fluorescence ( Fig. 7A and B) and a patchy cathodoluminescence ( Fig. 7C and D). Calcispheres of type C, in contrast, are filled with micrite and their walls are encrusted by blocky calcite (Fig. 6E and F). Raman spectroscopy revealed that the internal fillings of all calcisphere types consist of calcite, while their micritic walls additionally contain organic matter ( Fig. 8C and F).

Bulk geochemical characterization
The TIC and TOC contents are highly variable (  fig. 2). Carbonate contents vary between ca 11.3 wt.% and 91.9 wt.% (beds P 14 and P 1 , respectively), with an average value of 65.5 wt.%. TOC contents range  Programmed pyrolysis ('Rock-Eval') derived T max values vary between 440°C and 449°C (Table 1). HI and OI values range from 87 to 154 mg HC g À1 TOC and from 3 to 27 mg CO 2 g À1 TOC, respectively. The production index (PI) values range between 0.16 and 0.29 (Table 1; Pei, 2022).
Stable carbon isotopes (d 13 C carb and d 13 C org ) d 13 C carb values range between 0.7& and 4.2& (beds T 12 and P 1 , respectively), with an average value of 1.6&. d 13 C carb value is highest at the base of the section (4.2& in bed P 1 ) and then decreases significantly towards the top (1.1& in bed T 18 ). This trend is only interrupted by bed P 5 , which shows a higher d 13 C carb value of 2.4& (Table 1; Fig. 9; Pei et al., 2021, fig. 2). Table 1. Bulk geochemical data, including carbonate (wt.%), total organic carbon (TOC, wt.%), temperature at maximum yields of S2 hydrocarbons (T max ,°C), quantities of free hydrocarbons (S1, mg HC g À1 rock), hydrocarbons yielded from labile kerogen (S2, mg HC g À1 rock), the CO 2 generated from organic carbon at higher temperatures up to 650°C (S3, mg CO 2 g À1 rock), the hydrogen index (HI; S2/TOC 9 100), the oxygen index (OI; S3/TOC 9 100) and the production index (PI; S1/(S1 + S2)) (partially modified from Pei et al., 2021;Pei, 2022    Sedimentological and geochemical characteristics of the Lahnewies Syncline Section, Northern Calcareous Alps

Field observations and microfacies analyses
The combined Lahnewies Syncline section is about 60 m thick. No volcanic ash beds/layers were recorded, consistent with XRD data analyses (Table S1; Fig. 4). The section begins with the Hochalm Member, consisting of framestones characterized by corals and coralline sponges that are associated with green algae (Dasycladaceae) and large molluscs (Neomegalodon)  (Figs 4 and 10A). The lower bed is characterized by a mass occurrence of juvenile individuals of the ammonoid Choristoceras ( Fig. 10B; G€ umbel, 1861;Reitner, 1978). The upper bed contains the ammonoids Eopsiloceras planorboides and Choristoceras ammonitiforme ( Fig. 10C; G€ umbel, 1861). Furthermore, this bed also preserves a diverse bivalve community, dominated by the thin-shelled pectinid Agerchlamys. The ammonoid C. ammonitiforme indicates a stratigraphic position below the C. marshi zone (Wiedmann et al., 1979;Krystyn, 1987;Karl et al., 2014;Galbrun et al., 2020). The mass occurrence of relatively wellpreserved E. planorboides specimens in the upper bed is also notable because this phylloceratid ammonite is generally rare and restricted to the Lahnewies Syncline. The originally aragonitic shells of E. planorboides individuals are diagenetically dissolved and have not been replaced, whereas the periostracum is mostly preserved in its original form. If freshly exposed, E. planorboides typically exhibits a golden colour, while it appears white in specimens that are exposed to air. The Eopsiloceras Bed consists of mudstones to wackestones. In the upper part, the bed locally contains abundant framboidal pyrite and shows evidence for significant bioturbation by Zoophycos (Fig. 10D). The micrite mainly contains globular nannofossils such as Prinsiosphaera triassica (abundant) (Figs 11A, 11B and 12A to C) and Orthopithonella (subordinate), both of which were interpreted as dinoflagellate cysts (Jafar, 1983;Janofske, 1987Janofske, , 1992Demangel et al., 2020). Coccolith remains are also abundant but poorly preserved due to diagenesis ( Fig. 11A and B). However, most coccoliths could be assigned to Archaeozygodiscus, Crucirhabdus and Cleospharea (Janofske, 1987;Bown, 1987a,b;Bown et al., 2004). Occasionally, other nannoliths such as Conusphaera (Jafar, 1983;Janofske, 1987Janofske, , 1992 as well as spherical aggregates of calcite crystal plates occurred, which probably also represented mineralized haptophytes. Taken together, calcified planktonic algae are the major constituent of the carbonate sediment. A few metres above the Eopsiloceras Bed, the ca 50 to 100 cm thick T-Bed follows, which marks the top of the Eiberg Member (Figs 4 and  13) and consists of mudstones to wackestones ( Fig. 14A and B). Similar to the Eopsiloceras Bed, the T-Bed contains Zoophycos burrows, but these traces successively disappear towards the top part (Fig. 13E). The T-Bed also contains the ammonoid C. marshi in the top layer (Fig. 14C), which confirms a late Rhaetian age.
The Tiefengraben Member begins above the T-Bed and consists of black mudstones that contain only few microfossils (for example, the ostracod Bairdia sp., and the unusual foraminifer Praegubkinella sp.) but no nannofossils. In the lower part, the Tiefengraben Member locally exhibits organic-lean carbonate concretions ( Fig. 15A and B). The overlying Schattwald Beds are characterized by pink colour, carbonate-poor mudstones with an absence of fossils (Fig. 15C). The Breitenberg Member atop, in contrast, consists of wackestones and contains abundant fossils of demosponges, radiolarians and benthic foraminifera ( Fig. 15D and E). Particularly noteworthy is the ammonoid Psiloceras naumanni (Fig. 15F to H), which is indicative for the Calliphyllum zone, and hence a Hettangian age.

Lipid biomarkers
The aliphatic fractions of all studied TOC-rich samples contain abundant n-alkanes as well as acyclic head-to-tail linked isoprenoids, such as pristane and phytane (Fig. 17A). Furthermore, a clear odd-over-even predominance is detected between n-C 24 and n-C 32 (CPI = 1.7-2.7; Table 2). Samples R 5 and R 6 from the T-Bed contain remarkably high abundances of n-C 37-39 (Fig. 17B).
Stable carbon isotopes (d 13 C carb and d 13 C org ) d 13 C carb values vary from À4.6& to 1.8& (samples R 8 and R 3 , respectively) (Table 1; Fig. 4). The Eopsiloceras Bed and the T-Bed both exhibit d 13 C carb values of ca 1&, while the Tiefengraben Member is characterized by a significantly lower d 13 C carb value of À4.6& (sample R 8 ).
d 13 C org values range between À30.1& and À26.5& (samples R 3 and R 8 , respectively) ( Table 1; Fig. 4). Both the Eopsiloceras Bed and T-Bed exhibit low d 13 C org values of ca À30& (Table 1; Fig. 4), which is in the range of the well-documented negative d 13 C org excursion at the T-J boundary (Ruhl et al., 2010. In the Tiefengraben Member, d 13 C org value is higher (À26.5& in sample R 8 ) (Table 1; Fig. 4). Ecosystem changes through the critical intervals 1619

Organic matter preservation
Sedimentary organic records might be compromised by post-depositional processes, particularly through thermal degradation during burial (Peters et al., 2005;Mißbach et al., 2016). Therefore, the thermal maturity of organic matter has to be critically assessed prior to further geobiological interpretations (Love & Zumberge, 2021). Programmed pyrolysis data are powerful means for such assessments, but can only be considered robust if TOC is ≥0.2 wt.% and S2 is >0.2 mg HC g À1 rock (Peters, 1986;Peters & Cassa, 1994). All samples analysed herein meet these criteria and can be considered reliable (Table 1; Pei, 2022).
In the Lahnewies Syncline section, T max values (<430°C ; Table 1) indicate thermally immature organic matter. HI and OI values (>435 mg HC g À1 TOC and <35 mg CO 2 g À1 TOC, respectively; Table 1) are indicative for type II kerogens that typically form in marine environments (Peters, 1986;Peters & Cassa, 1994). The low thermal maturity of organic matter is in good accordance with PI <0.1 and CPIs >1 (Tables 1  and 2). Because of the low thermal maturity of

Primary producers through the P-T and T-J critical intervals
In the P-T boundary section in Xiakou, masses of calcispheres occur in one distinct carbonate layer P 15-2 (MF-5) that is interlayered between volcanic ash interlayers (layers P 15-1 and P 15-3 ) and characterized by peloidal textures (Fig. 6A and B). In the T-J boundary section, calcispheres are also quite abundant. The term 'calcisphere' derives from Calcisphaera, a genus originally established for specific calcareous microfossils in Carboniferous carbonates (Williamson, 1880), but is nowadays commonly used for spherical calcareous microfossils of various origins.
Calcispheres in Upper Triassic or younger rocks were commonly interpreted as cysts of calcareous dinoflagellates (Keupp, 1991;Wendler & Bown, 2013), perhaps reflecting the onset of significant calcification of pelagic plankton . The biological affinities of calcispheres in Middle Triassic and older rocks, in contrast, are disputed. For instance, it has been proposed that some Silurian calcispheres might represent ancestors of dinoflagellates (Servais et al., 2009), whilst specimens from the Late Devonian were interpreted as radiolarians (Antoshkina, 2006). Abundant calcispheres in microbialites at the P-T boundary (e.g. Pei et al., 2019;Zhang et al., 2020) perhaps represented coccoid cyanobacteria . In the light of these findings, the biological affinity of calcispheres in layer P 15-2 at Xiakou (Figs 6C to 6F, 7 and 8) remains unclear, although it is tempting to speculate that they represent ancestors of dinoflagellates.
Remarkably, n-alkanes in samples from the T-J boundary section commonly show a marked increase in abundance between C 37 and C 39 , particularly in samples R 5 and R 6 (Fig. 17B). This distinct feature requires explanation as the relative abundance of n-alkanes >n-C 33 usually drops with increasing carbon number. N-alkanes with carbon numbers between 37 and 39 might derive from long-chain alkenones, which are ubiquitous in recent haptophytes (Cranwell, 1985;Brassell et al., 1986;Prahl et al., 1988). Haptophytes became abundant in the Triassic (De Vargas et al., 2007), but to the best of present knowledge alkenones have not been found in rocks older than the Jurassic (Dumitrescu & Brassell, 2003). However, it was experimentally demonstrated that the mild reduction of alkenones, as to be expected during diagenesis, results in the formation of n-alkanes with identical chain lengths (Love et al., 2005), supporting a precursor-product relationship between these compounds. It thus appears plausible that the observed n-C 37-39 alkanes are diagenetic derivatives of alkenones that originated from early haptophytes such as coccolithophorids.
Ecosystem changes through the P-T and T-J critical intervals -Is volcanism the common driver?
The P-T and T-J crises both have been intimately associated with large continental flood basalt provinces (i.e. the Siberian Traps and CAMP, respectively) (e.g. Whiteside et al., 2010;Blackburn et al., 2013;Burgess & Bowring, 2015;Thibodeau et al., 2016;Burgess et al., 2017;Davies et al., 2017;Percival et al., 2017;Broadley et al., 2018;Wang et al., 2018). The Siberian Traps are one of the largest known continental flood basalt provinces of the Phanerozoic, with an estimated original volume of ca 3 9 10 6 km 3 (Reichow et al., 2002). About two-thirds of the total lava/pyroclastic volume erupted over a period of only ca 300 000 years, shortly before and concurrent with the end-Permian mass extinction (Burgess & Bowring, 2015), and concurrent with a massive emplacement of sill intrusions (Burgess et al., 2017). Simultaneously, the volcanism liberated significant amounts of volatiles such as halogens, carbon dioxide, sulphur dioxide and methane (Svensen et al., 2009;Ogden & Sleep, 2012;Broadley et al., 2018;Clapham & Renne, 2019), perhaps resulting in a six-fold increase of atmospheric pCO2 during this crisis (Wu et al., 2021).
Negative excursions of both d 44/40 Ca (in carbonates and hydroxyapatites) and d 13 C (in carbonates and organic matter) indicated that the elevated pCO2 levels may have caused ocean acidification (Payne et al., 2010;Hinojosa et al., 2012;Silva-Tamayo et al., 2018). Indeed, shallow marine successions from that time commonly exhibit an erosional surface between latest Permian carbonates and post-extinction microbialites, perhaps reflecting a decreased seawater pH Collin et al., 2009;Lehrmann et al., 2015;Pei et al., 2019). Acidification, together with physiological effects such as hypercapnia, seems to be consistent with the selective extinction of organisms with calcareous shells and poorly buffered respiratory physiology (Knoll et al., 2007;Clapham & Payne, 2011; for an alternative view see Slater et al., 2022).
At Xiakou, the calcisphere-rich layer P 15-2 (MF-5) is intercalated in volcanic ash (layers P 15-1 and P 15-3 ) (Fig. 2), perhaps indicating a direct relationship between volcanism and planktonic primary production. One plausible explanation is that high nutrition associated with volcanism resulted in a fertilization of marine environments, thereby promoting primary production in the ecosystems. This hypothesis is supported by studies that demonstrated a positive impact of volcanic eruptions on primary production in modern marine environments such as the Korean coast and the western Pacific Ocean (Zhang et al., 2017;Kim, 2020).
Like the Siberian Traps, CAMP classifies as one of the largest known Phanerozoic continental flood basalt provinces (Kravchinsky, 2012). The formation of the CAMP and large-scale intrusive processes are synchronous to the T-J extinction event, possibly implying a direct connection (Blackburn et al., 2013;Davies et al., 2017). Marine sedimentary successions show negative d 13 C carb and d 13 C org excursions at the time of CAMP (Ruhl et al., 2010Whiteside et al., 2010;Corso et al., 2014). Furthermore, a biocalcification crisis occurred during this time, which was likely linked to ocean water acidification because organisms with aragonitic or high-Mg calcite skeletons and a weak physiological control on biomineralization were particularly affected (Hautmann et al., 2008;Ruhl et al., 2010).
The T-J Lahnewies Syncline section records three regionally widespread anoxic events. The first two events are characterized by mass extinctions of certain ammonoids, as expressed in the names of respective beds (Choristoceras Bed and Eopsiloceras Bed) ( Fig. 4; Reitner, 1978;Karl et al., 2014). The third event is reflected by the top layer of the T-Bed and almost results in a collapse of carbonate production. Palaeontological and organic geochemical data evidence a eutrophic ecosystem dominated by algae such as calcified haptophytes (i.e. coccolithophores) and dinoflagellates during the event (Figs 11C, 11D, 12D to 12M and 17;McRoberts et al., 2012;Hillebrandt et al., 2013). Notably, biomarkers indicative for water column stratification and photic zone anoxia (for example, gammacerane, isorenieratane), as detected in contemporaneous sections from other regions (Richoz et al., 2012;Blumenberg et al., 2016), have not been found.
The Lahnewies Syncline section provides no evidence for a direct influence of CAMP-related volcanic activity on uppermost Rhaetian environments. For instance, clay minerals that are commonly sourced by volcanic eruptions (for example, montmorillonite, bentonite, possibly smectite) have not been observed. The only potential indication is clay spherules contained in the topmost K€ ossen Formation of the Kendelbach section, which perhaps represents altered volcanic ash layers (Zajzon et al., 2012). However, CAMP might have affected local environments by supplying high quantities of fertilizing volatiles, carbon dioxide and sulphur dioxide to the atmosphere, resulting in acidic rain, and leading to ocean eutrophication (Davies et al., 2017). Thus, a possible relationship between volcanism and abundant dinoflagellates and coccolithophorids in the top layer of the T-Bed is plausible and requires further investigation (Figs 11C,11D,12D to 12M and 17).

CONCLUSIONS
This study focuses on ecological upheavals through the Permian-Triassic (P-T) and Triassic-Jurassic (T-J) critical intervals, which are commonly assumed to be caused by voluminous volcanic provinces [Siberian Traps and Central Atlantic Magmatic Province (CAMP), respectively]. Calcispheres, interpreted as dinoflagellates and/or coccolithophorids, are abundant in deposits from both intervals. Additionally, the T-J section preserves abundant 4-methyl sterenes (diagnostic for dinoflagellates) and C 37-39 nalkanes (indicative for haptophytes). In the P-T boundary section from South China, calcispheres occur in a distinct limestone bed that is intercalated in thick volcanic ash beds, which is probably related to the Siberian Traps. Thus, volcanism might have had a positive effect on the planktonic primary producers, perhaps by delivering essential nutrients. The studied T-J boundary sections from Northern Calcareous Alps, in contrast, preserve no sedimentological evidence for volcanic activity. Thus, a possible relationship between volcanism and blooms of dinoflagellates and coccolithophorids remains elusive and needs further investigation.
Hause-Reitner, D. Kohl, G. Scheeder, J. Dyckmans, J. Sch€ onig, K. L€ unsdorf, K. Wemmer, P. Adam, T. Wasselin and W. Dr€ ose are acknowledged for lab. assistance. This study was financially supported by the China Council Scholarship (CSC) and a Teach@T€ ubingen Fellowship by the University of T€ ubingen (Germany). The Forestry Department Oberammergau (Bayerische Staatsforsten) gave permission to drive on closed forest roads. Open Access funding enabled and organized by Projekt DEAL.