- The Paleontological Society
High-resolution palynological data sets from shallow marine Triassic-Jurassic (Tr/J) boundary beds of two principal sections in Europe (Hochalplgraben in Austria and St. Audrie's Bay in the United Kingdom) were analyzed to reconstruct changes in vegetation, biodiversity, and climate. In Hochalplgraben, a hardwood gymnosperm forest with conifers and seed ferns is replaced by vegetation with dominant ferns, club mosses and liverworts, which concurs with an increased diversification of spore types during the latest Rhaetian. Multivariate statistical analysis reveals a trend to warmer and wetter conditions across the Tr/J boundary in Hochalplgraben. The vegetation changes in St. Audrie's Bay are markedly different. Here, a mixed gymnosperm forest is replaced by monotonous vegetation consisting mainly of Cheirolepidiaceae (80–100%). This change is caused by a transition to a warmer and more arid climate. The observed diversity decrease in St. Audrie's Bay affirms this interpretation. Although both sections show major vegetation changes, neither of them demonstrates a distinctive floral mass extinction. A compilation of Tr/J boundary sections across the world demonstrates the presence of Cheirolepidiaceae-dominated forests in the Pangaean interior and increases in abundance of spore-producing plants adjacent to the Tethys Ocean. We propose that the non-uniform vegetation changes reflected in the Tr/J palynological records are the result of environmental changes caused by Central Atlantic Magmatic Province volcanism. The increase in greenhouse gases caused a warmer climate and an enhanced thermal contrast between the continent and the seas. Consequently, the monsoon system got stronger and induced a drier continental interior and more intensive rainfall near the margins of the Tethys Ocean.
The transition from the Triassic to the Jurassic Period, ca. 201.58 Ma (Schaltegger et al. 2008), is characterized by a major biotic crisis in the marine and terrestrial realms (e.g., Hallam 2002; Olsen et al. 2002; Tanner et al. 2004). However, the severity and patterns (i.e., abrupt, stepwise, or gradual) of this crisis, especially for the terrestrial realm, are not well understood (Hallam 2002; Bambach et al. 2004; Lucas and Tanner 2008). Explanations for the biotic turnover have included both gradualistic and catastrophic mechanisms (e.g., Hallam and Wignall 1997; Tanner et al. 2004; Hesselbo et al. 2007). A frequently proposed mechanism is massive volcanism of the Central Atlantic Magmatic Province (CAMP), one of the largest known flood basalt provinces, related to the breakup of Pangaea (Wignall 2001; Hesselbo et al. 2002; Knight et al. 2004; Marzoli et al. 1999, 2004; Schoene et al. 2010; Deenen et al. 2010). The release of large amounts of carbon dioxide (CO2) and other toxic gases in the atmosphere by CAMP volcanism induced climate change and could have caused biotic disturbance (McElwain et al. 1999; Hesselbo et al. 2002; Tanner et al. 2004; Van de Schootbrugge et al. 2009). However, new compound-specific C-isotope records of long chain n-alkanes show a ∼8.5‰ negative excursion, which cannot be explained by the release of volcanic CO2 (Ruhl et al. 2011). These data therefore point to a short-term injection of 12 × 103 gigatons of isotopically depleted carbon as methane from clathrates into the atmosphere. The injection of methane from melting clathrates may have been triggered by the beginning CAMP volcanism. An alternative catastrophic mechanism causing the extinction is an extraterrestrial impact (Olsen et al. 2002), but evidence for such a scenario is rather limited.
Palynological analysis is a useful method in unraveling past vegetation patterns and reconstructing climate change (e.g., Barrón et al. 2006; Galfetti et al. 2007). Palynological records across the Triassic/Jurassic (Tr/J) transition, however, are controversially discussed because of the paucity of sections with a sufficient time resolution and/or well-established stratigraphic framework. Furthermore, many records are only qualitative or semiquantitative (i.e., absence/presence data). Evidence for a prominent end-Triassic extinction event in the plant fossil record is ambiguous. A major extinction of 60% of sporomorph taxa followed by a sharp spore spike has been shown in the Newark Basin in the United States (Fowell and Olsen 1993; Fowell et al. 1994; Olsen et al. 2002). A recent study from the Germanic Basin showed a severe vegetation shift from gymnosperm forests to pioneer assemblages across the Tr/J boundary, linked to CAMP volcanism (Van de Schootbrugge et al. 2009). Other palynological studies from Europe show less severe and more gradual changes in assemblages at the Tr/J transition (e.g., Warrington 1974; Morbey 1975; Lund 1977; Schuurman 1979; Achilles 1981; Kürschner et al. 2007; Bonis et al. 2009). Although palynological data sets give information about the regional vegetation changes, macrofossil records give an indication of the local vegetation. Quantitative macrobotanical data from East Greenland showed that Triassic forests with high-diversity communities were replaced by lower-diversity forests and that there was a gradual extinction prior to the Tr/J boundary (McElwain and Punyasena 2007; McElwain et al. 2007). Although the Late Triassic event in Greenland did not induce mass extinction of plant families, it is accompanied by major and abrupt changes in floral ecology and diversity (McElwain et al. 2009). The palynological record from Greenland shows a different pattern, with no major diversity or assemblage changes and no conclusive evidence for an extinction event (Raunsgaard Pedersen and Lund 1980; Koppelhus 1997). The discrepancy between local macro- and microfloral records has been ascribed to differences in reproductive biology between gymnosperms groups, resulting in differences in pollen and/or spore output (Mander et al. 2010). For example, cycads and Bennettitales appear to be underrepresented in the pollen assemblages in comparison to the macrofossil record because they are thought to be insect pollinated.
This paper is aimed at a reconstruction of vegetation history, diversity patterns, and inferred climate changes across the Tr/J transition on the basis of high-resolution quantitative palynological data sets from paleogeographically contrasting settings in Europe. Comparison of records from Hochalplgraben (Austria) and St. Audrie's Bay (United Kingdom) with each other and with various boundary sections across the world may contribute to better understanding of the nature and rate of Tr/J vegetation changes, as well as the environmental or climate changes that could have driven this turnover.
Materials and Methods
The present study concentrates on two European principal Tr/J boundary sections (Fig. 1): Hochalplgraben in the Northern Calcareous Alps in Austria (47°28′20′′N, 11°24′42′′E) and St. Audrie's Bay in Southwest U.K. (51°11′N, 3°17′W). A description of the Hochalplgraben section, together with a high-resolution palynological study is presented by Bonis et al. (2009). Hochalplgraben's palynological assemblages are similar to those of the Tiefengraben section (Kürschner et al. 2007) and the Kuhjoch section (Bonis et al. 2009). The latter is recently approved as the Global Stratotype Section and Point (GSSP) for the base of the Jurassic (Von Hillebrandt et al. 2007). St. Audrie's Bay is a classic outer-Alpine Tr/J boundary section (Hounslow et al. 2004; Warrington et al. 1994, 2008). Investigated samples for a detailed palynological study come from the Westbury Formation, the Lilstock Formation (Cotham Member and Langport Member), and the lower part of the Blue Lias Formation. The first occurrence (FO) of the Psiloceras spelae tirolicum ammonite has been chosen as the primary boundary marker for the base of the Jurassic period (Von Hillebrandt et al. 2007; Von Hillebrandt and Krystyn 2009). Because this ammonite is absent in St. Audrie's Bay, the Tr/J boundary is indicated as a dashed line at the appearance of the Jurassic ammonite Psiloceras planorbis. The FO of Cerebropollenites thiergartii is approximately contemporaneous with the FO of P. spelae tirolicum (Bonis et al. 2009; Kürschner and Herngreen 2010). This pollen species is present in both sections and it is therefore a useful palynological marker for the base of the Jurassic. The onset of the CAMP volcanism is linked to the initial negative carbon isotope excursion and pre-dates the Tr/J boundary by ∼100,000 years (Deenen et al. 2010; Ruhl et al. 2010).
The majority of the pollen and spores of land plants (sporomorphs) from the Hochalplgraben and St. Audrie's Bay sections can be classified (Appendix 1) in terms of their botanical affinity (Schulz 1967; Balme 1995; Abbink 1998; Hubbard and Boulter 2000; Herngreen 2005a,b; Raine et al. 2005; Barrón et al. 2006; Lindström and Erlström 2006; Ziaja 2006; Traverse 2007; Van Konijnenburg-Van Cittert personal communication 2009). We used these affinities to reconstruct vegetation patterns across the Tr/J boundary. It has been demonstrated in studies on modern palynology- vegetation relationships that modern pollen diversity reflects the diversity of the surrounding vegetation and that fossil pollen diversity may provide an important proxy in reconstructing changes in the plant diversity (e.g., Weng et al. 2007; Lézine et al. 2009; Pelánková and Chytrý 2009). However, dealing with Mesozoic material, one factor that is difficult to include is the variation in sporomorph production by the parent plants. For example, because Cheirolepidiaceae were most probably wind pollinators, they produced a large amount of pollen (Classopollis) per individual tree (Alvin 1982; Ziaja 2006). However, for the majority of Triassic and Jurassic plant species the pollination strategy is not known and can only be inferred from the overall appearance of the plant and the pollen morphology. For some sporomorph taxa the botanical affinity is even still unknown because the sporomorphs have never been found in situ (e.g., Trachysporites spp. and Ovalipollis spp.). This should be kept in mind when interpreting changes in vegetation composition based on sporomorph counts. Furthermore, the relative abundance of pollen and spores can be influenced by sea level change (e.g., Abbink et al. 2004). During high sea level one would expect an increase of pollen with a high buoyancy (e.g., bisaccates). Spores would be more abundant at low sea level because they are relatively heavier and more difficult to transport. This is known as the Neves effect (e.g., Traverse 2007).
Multivariate Statistical Analysis
A detrended correspondence analysis (DCA) was carried out on the relative abundances of sporomorphs (with a square root transformation of species data and down-weighting of rare species) to determine the gradient length of the first axis. The gradient length of the first axis is 2.026 standard deviations (SD) for Hochalplgraben and 2.393 SD for St. Audrie's Bay. Because the gradient lengths of the data sets did not exceed 3 SD, a linear ordination method, Principal Components Analysis (PCA), was used to make a summary of the relative pollen and spore abundances data sets (Lepš and Šmilauer 2003). Both PCAs were done with a square-root transformation of the species data and the data were centered by variables (taxa).
We used the qualitative sporomorph data sets from Hochalplgraben and St. Audrie's Bay to carry out a diversity analysis with the computer program PAST (Hammer et al. 2001). Data were subjected to the range-through assumption (absences between first and last appearance are treated as presences) and possibly reworked taxa were rejected from the diversity analysis. Additionally, we calculated the amount of pollen and spore taxa present per sample (species richness). The quantitative data sets were used for a rarefaction analysis (Birks and Line 1992) standardized on a pollen sum of 207 grains with 95% confidence intervals. This is an intrapolation technique making it possible to estimate how many species would have been found had the sample been smaller than it actually was (Raup 1975). In this way we can compare estimated diversities at a constant sample size.
In the latest Triassic Kössen Formation in the Hochalplgraben section the vegetation can be described as a hardwood gymnosperm forest with conifers and seed ferns (Fig. 2). The most dominant conifers are Cheirolepidiaceae, the parent plants of Classopollis pollen. At the transition from the Kössen Formation to the Tiefengraben Formation spore-producing plants increase considerably. Most abundant are liverworts, club mosses, and different fern types like Dicksoniaceae and Cyatheaceae (both tree ferns), Schizaeaceae (climbing ferns), Matoniaceae, and Osmundaceae. Cheirolepidiaceae are decreasing whereas Caytoniales (seed ferns) are increasing. A narrow peak (>90%) of Cheirolepidiaceae is present at 550 cm. Above the Schattwald beds, Cheirolepidiaceae decrease and the seed ferns almost disappear. Abundant vegetation groups are different fern types and Selaginellales (spike mosses). “Trachysporites producing ferns” constitute a major part within the “other ferns” group. Above 800 cm (the earliest Jurassic), the vegetation is relatively stable, consisting mainly of liverworts and ferns. Gymnosperms are no longer an abundant element of the vegetation and the relative amount of pollen of Cheirolepidiaceae decreases to values of around 10%.
The vegetation changes in the St. Audrie's Bay section differ from the observed change in the Hochalplgraben section (Fig. 3). Up to the top of the Westbury Formation major components are Cheirolepidiaceae, other gymnosperms, and liverworts. The lower part of the Lilstock Formation (Cotham Member) consists of Caytoniales, Taxodiaceae, and an alternating amount of Cheirolepidiaceae and different fern types (e.g., Schizaeaceae, Matoniaceae, Osmundaceae). At ∼1250 cm horsetails increase in abundance. The most striking feature in the upper Lilstock Formation (Langport Member) is the acme (>80%) of club mosses (Selaginellales). Alternating fern and Cheirolepidiaceae abundance make up the rest of the Langport Member. The Blue Lias Formation shows a monotonous vegetation consisting of Cheirolepidiaceae (80–100%) combined with club mosses (Selaginellales).
Multivariate Statistical Analysis
The results from Hochalplgraben and St. Audrie's Bay are displayed as the species scores on the first and second axes of PCA ordination diagrams (Fig. 4). These two main ordination axes are the dimensions through the data set that explain the largest variance in species composition and can be translated in terms of the environmental and/or climatic gradient that controls the data set.
The first axis of the Hochalplgraben data set explains 43.5% of the total variance within the data set, and the second axis explains 19.7% (Fig. 4A). On the positive side of the first axis, various pollen taxa have a high score (e.g., Classopollis meyeriana, Classopollis torosus, Vitreisporites pallidus + bjuvensis, and Ovalipollis pseudoalatus). Spores like Trachysporites fuscus, Ricciisporites tuberculatus, and Heliosporites reissingeri have a high negative score on the first axis. These spores are produced by moisture-loving plants as ferns and liverworts. Cheirolepidiaceae, which produced Classopollis pollen, became dominant or even mono-dominant during increasing aridification, which limits the areal extent of the moisture-loving plants (Vakhrameev 1987). They preferred a subtropical to tropical, somewhat arid climate (Vakhrameev 1981, 1991), although there are some species that may have had a coastal habitat (e.g., Batten 1974; Watson 1988; Abbink 1998). However, from the position of the spores relative to the position of Classopollis, the first axis can then be interpreted as a ratio between sporomorph types indicative of relatively wet and relatively dry conditions. C. meyeriana also has a very high score on the positive side of the second axis. Taxa with a negative score on the second axis are C. torosus, R. tuberculatus, Vitreisporites pallidus + bjuvensis, Convolutispora microrugulata, and Deltoidospora spp. We considered cheirolepidiaceous conifers to be thermophilous (Vakhrameev 1981; Alvin 1982). Studies from Greenland (Raunsgaard Pedersen and Lund 1980; Koppelhus 1997) and Siberia (Rovnina 1972) report that Classopollis is rare, and if it occurs in the record, it is mostly C. torosus. The occurrence of mainly C. torosus in high-latitude records suggests that this pollen type was produced by a Cheirolepidiaceae species better adapted to cooler conditions than the parent plant of C. meyeriana. Therefore, the second axis is interpreted to represent a ratio between sporomorph types indicative of relatively cooler versus relatively warm conditions. The high negative score on the second axis of Vitreisporites, a taxon also known from Greenland (Raunsgaard Pedersen and Lund 1980), confirms this interpretation.
The first and second PCA axes from St. Audrie's Bay, respectively, 46.5% and 17.8% of the total variance within the data set (Fig. 4B). The position of the taxa in the ordination plot is similar to Hochalplgraben. To facilitate a direct comparison, the axes from St. Audrie's Bay are reflected (Fig. 4B). C. meyeriana has a high negative score on the first axis, while C. torosus, O. pseudoalatus, Rhaetipollis germanicus and R. tuberculatus score high on the positive side of the first axis. On the second axis, spore taxa like H. reissingeri, Deltoidospora spp., Conbaculatisporites spp., Trachysporites spp., Todisporites minor + major and Acanthotriletes varius have positive scores, whereas various pollen taxa (e.g., C. meyeriana, C. torosus, O. pseudoalatus, R. germanicus, and Granuloperculatipollis rudis) have high negative scores. The first axis is interpreted as a relative temperature axis and the second as a humidity axis using the same way of reasoning as discussed above for Hochalplgraben. We recognize that additional or other environmental factors may have contributed to axes 1 and 2 scores; however, the most parsimonious explanation based on the available data is that axes 1 and 2 represent temperature and humidity, respectively.
All three diversity records (range-through, species richness, and rarefaction) show a similar trend for both sections (Fig. 5). There is an almost fourfold diversity increase in Hochalplgraben from the Eiberg Member to the top of the Schattwald beds (Fig. 5A). The increase in diversity coincides with the initial negative CIE. Highest diversity is reached in the Schattwald beds with 80–90 taxa in the range-through record. A sharp decline in species richness occurs at the top of the Schattwald beds, which is also present in the diversity curve based on rarefaction. The sudden jump in diversity is in part due to asedimentary hiatus caused by a fault in this section at this level (for further discussion see Bonis et al. 2009). Higher up in this section the pollen and spore diversity remain rather stable.
In St. Audrie's Bay, the diversity increases slightly from the base of the Westbury Formation and decreases from the upper part of the Westbury Formation into the Blue Lias Formation (Fig. 5B). The highest diversity in the Westbury Formation is reached with 61 taxa. There are no marked changes in diversity across the level of the initial negative CIE.
It appears that no significant loss in sporomorph diversity occurs at the level of the initial negative isotope excursion, which is contemporaneous with the main phase of the end-Triassic marine mass extinction. Instead, the Hochalplgraben section shows a marked increase in diversity. The increasing and decreasing diversity at the base and top of the records is probably affected by the edge effect, which is an exaggerated high concentration of first occurrences at the beginning and a high concentration of last occurrences at the end of a record (Boltovskoy 1988; Jaramillo 2002). However, the parallel trend with the species richness record implies that the range-through record shows true changes in the diversity.
Climate Changes in Hochalplgraben and St. Audrie's Bay
The PCA ordination diagrams of Hochalplgraben and St. Audrie's Bay show comparable distributions of taxa (Fig. 4). This is an important result, because it implies that despite the different patterns of vegetation changes in both sections, the sporomorph assemblages are similar in composition. In Hochalplgraben, axis 1 is the humidity axis and axis 2 the temperature axis, whereas this is the opposite in St. Audrie's Bay. This is explained by the observation that the overall shift in vegetation composition in Hochalplgraben is from Cheirolepidiaceae to spore-producing plants (Fig. 2), probably related to the change in humidity levels (e.g., Vakhrameev 1981, 1987, 1991). Such an overall shift between Cheirolepidiaceae and spore-producing plants is not present in St. Audrie's Bay (Fig. 3). Instead, there are changes between different conifer species where relative temperature might be a more important factor.
The sample scores on the first and the second axes from the PCA plotted through time result in relative humidity and temperature records for Hochalplgraben (Fig. 6A) and St. Audrie's Bay (Fig. 6B). The initial negative carbon isotope excursion (CIE) (e.g., Hesselbo et al. 2002; Ruhl et al. 2009) and the first occurrence (FO) of Cerebropollenites thiergartii are used as correlation lines. A minor hiatus is present in the lower part of the Tiefengraben Member (Bonis et al. 2009) from Hochalplgraben. This corresponds to the interval between 1450 and 1850 cm in St. Audrie's Bay (Fig. 6B, gray interval in δ13C curve). Our correlation between both realms is in agreement with Hallam's (1981) suggestion based on Schuurman's (1977, 1979) data: that the Zlambach Beds, most of the Kössen Beds, and most of the Westbury Formation (together with the topmost part of the underlying Mercia Mudstone Group) are age-equivalent, whereas the topmost Kössen Beds correspond to part of the Lilstock Formation up to the upper part of the Cotham Member. The Schattwald beds correspond to the Langport Member of the Lilstock Formation.
Figure 6 shows the relative changes in temperature and humidity inferred from the sample scores on the first and second axis. Prior to the initial CIE the vegetation was dominated by conifers (Figs. 2, 3) and the climate was dry and warm in the western Tethys realm and humid and warm-temperate in the northwestern European realm. It must be noted that the temperature gradient reconstructed by the PCA was probably small because a warm climate predominated the Tr/J boundary interval (e.g., Frakes 1979; Sellwood and Valdes 2007). The initial CIE coincides with peak temperatures and after a brief dry episode a wetter period. The abrupt thermal maximum during the initial CIE is likely caused by methane release from melting clathrates (Ruhl et al. 2011). This interval is followed by cooling and increasing humidity in both sections. A palynofloral transition interval with four pronounced spore peaks is present in the Lilstock Formation in St. Audrie's Bay (Fig. 3). Instead of interpreting this interval as a pioneer vegetation caused by CAMP volcanism, we think these spore peaks are related to CAMP induced climate changes, causing a stronger seasonality and changes in the strength of the monsoon (Bonis et al. 2010a). In the Tiefengraben Member (from 550 cm) in Hochalplgraben and in the Blue Lias Formation (from 1850 cm) in St. Audrie's Bay, vegetation and climate patterns are completely different. Although warming is visible in both sections, Hochalplgraben is characterized by an increasingly wetter climate while drier climate prevails in St. Audrie's Bay (Fig. 6). This is visible in the vegetation diagram as the transition from a conifer-dominated assemblage to a vegetation type consisting mainly of ferns and liverworts in Hochalplgraben. By contrast, assemblages in St. Audrie's Bay were completely dominated by Cheirolepidiaceae during this time. Distinct changes in high-resolution δ18O records of fossil oysters from the United Kingdom (Korte et al. 2009) correspond to our reconstructed temperature axis from St. Audrie's Bay. Both proxy records show a warming trend from the Triassic to the Jurassic, interrupted by a cooler period. The fossil oyster record is not influenced by sea level changes. The similar climate signals from the O-isotope and the sporomorph records suggest that the changes in sporomorph assemblages reflect true climate induced vegetation changes and that the sporomorph record is not biased by sea level changes. Therefore, the sporomorph record is a useful proxy for climate change (i.e., temperature, humidity).
Diversity Patterns in Hochalplgraben and St. Audrie's Bay
The first important result is that the pattern is similar for all diversity methods (qualitative: range through and species richness, and quantitative: rarefaction) (Fig. 5). This implies that in this case even a relatively low sporomorph sum (∼200) is sufficient to get a general idea about diversity trends. However, for a better estimation of the number of taxa per sample we prefer diversity estimates based on qualitative pollen analysis, because the number of taxa can be two to three times higher (Fig. 5). In this case, moreover, plant taxa that are a rare but frequently occurring floral component are included in the diversity analysis. Besides true first or last occurrences, taxa can also be temporarily absent because of changing environmental conditions. Therefore, we have plotted the number of taxa per sample (species richness) next to the range-through diversity method. The two methods show parallel trends. An almost fourfold increase in pollen and spore diversity occurs across the Tr/J boundary (Fig. 5A) in Hochalplgraben, which is linked to a climate change from relatively hot and arid to more humid conditions (Fig. 6A). In contrast to Hochalplgraben, the St. Audrie's Bay section shows a drastic decrease in diversity from the latest Triassic to the earliest Jurassic in (Fig. 5B), caused by a change to a warm an arid climate (Fig. 6B). Whereas diversity increases to 80–90 taxa in the Schattwald beds, diversity in the time-equivalent beds from St. Audrie's Bay is much lower, about 50 taxa (Fig. 5). This can be explained by the different paleogeographic positions of the sections: Hochalplgraben was situated in a tropical summer-wet biome whereas St. Audrie's Bay was located in a warm temperate biome (e.g., Rees et al. 2000; Willis and McElwain 2002).
Figure 7 shows the diversity, extinction, and origination in Hochalplgraben and St. Audrie's Bay separately for the pollen and spores. The originations at the beginning and the extinctions at the end of the records are partly influenced by the edge effect. However, true disappearances in the end of the St. Audrie's Bay section (Fig. 7F) are observed as C. meyeriana becomes the monodominant species. The most striking character in the Hochalplgraben record is the gradual diversification of spores above the Kössen Formation above the initial negative CIE and the maximum diversity in the Schattwald beds (Fig. 7A) during the positive rebound interval. However, most of these spores are known to have much longer stratigraphic ranges than recorded in these sections and were already present earlier in the Rhaetian (Bonis et al. 2009; Kürschner and Herngreen 2010). The virtual lack of spores in the Kössen Formation is probably caused by non-favorable environmental conditions limiting spore-producing plant growth during deposition (Bonis et al. 2009): relatively arid and warm (Fig. 6A). The sudden diversity decrease at 550 cm in Hochalplgraben is not an extinction event, because a hiatus disturbs the record. There are no evident extinction horizons present in the Hochalplgraben or the St. Audrie's Bay section (Fig. 7). Some Late Triassic taxa in St. Audrie's Bay that disappear coincident with the initial negative CIE are Rhaetipollis germanicus, Ovalipollis pseudoalatus, and Lunatisporites rhaeticus. A possible explanation is that the mother-plants of these pollen types could not cope with an abrupt change to a warmer and/or wetter climate (Fig. 6B) (Ruhl et al. 2011).
The diversity increase in Hochalplgraben is in sharp contrast to floral diversity data from other regions. The contrasting trends in floral diversity likely reflect regional differences in environmental stress, climatic changes, and different paleogeographic positions during the end-Triassic biotic crisis. In the Eocene, a similar increase in spore abundance and diversity reflects increased rainfall in a tropical climate (Jaramillo 2002). This might be a Cenozoic analogy of the changes we see in Hochalplgraben (Figs. 5A, 6A), a site likely located in the tropical summer-wet biome during deposition of the Tiefengraben Member. There is a clear palynological turnover and a marked decline in sporomorph diversity of about 60% in the Passaic Formation below the Jacksonwald Basalt (Fowell and Olsen 1993). The significance of this turnover has been questioned, because this event pre-dates the Tr/J transition and the sedimentary succession is interrupted by a hiatus at this level (Van Veen 1995; Kozur and Weems 2005; Cirilli et al. 2009). A quantitative macrobotanical study from Greenland shows a decrease in standing species richness by about 85% (McElwain et al. 2007). Although the late Triassic event in East Greenland did not induce mass extinction of plant families, it accompanied major and abrupt change in their ecology and diversity (McElwain et al. 2009). Previous palynological analysis of the Greenland Tr/J transition did not show pronounced assemblage changes or an extinction event (Raunsgaard Pedersen and Lund 1980; Koppelhus 1997). It should be noted that taphonomic and biological factors, such as strategies for plant growth and reproduction, influence both the macro- and microfloral records and may explain part of the difference in diversity changes between these records. Sheet splay or crevasse splay deposits are the primary facies in which the plant fossils occur in the Astartekløft section (McElwain et al. 2007), implying a mix of autochthonous and allochthonous vegetation. On the other hand, abandoned channels would be expected to preserve predominantly parautochthonous plant communities growing in close proximity to the channel (McElwain et al. 2007). A detailed integrated macro- and microfloral study of the Astartekløft section revealed that differences between the macro and sporomorph records particularly occurs among reproductively specialized plants such as cycads and bennettites (Mander et al. 2010).
A comparison of the spore intervals from the Newark Basin (Fowell and Olsen 1993), Germanic Basin (Van de Schootbrugge et al. 2009), and St. Audrie's Bay (Bonis et al. 2010a) shows that there is no single unambiguous global end-Triassic spore spike. Several short-term fluctuations in spore abundance were observed in the St. Audrie's Bay palynomorph records, reinforcing the idea that the spore peaks are related to climate changes, i.e. changes in monsoon strength and in precipitation and/or humidity (Bonis et al. 2010a). The diversity based on a range chart compiled from several Triassic studies and reviews also shows only a minor decline in sporomorph diversity (∼20%) across the Tr/J transition (Kürschner and Herngreen 2010). This is mainly the result of a long-term decrease in the numbers of spore species throughout the Rhaetian (Kürschner and Herngreen 2010). Furthermore, it should be taken into account that this Tr/J diversity curve concerns stages and that typical early Rhaetian sporomorphs such as vesicate and bisaccate pollen are included in the diversity calculation. These are not included in our study, which includes only the latest Rhaetian and earliest Hettangian.
To summarize, we suggest that there is only a minor qualitative palynological extinction event across the Tr/J boundary, limited to the interior of Pangaea. Instead, the transition interval is characterized by climate-induced major quantitative changes in the sporomorph assemblages.
Global Triassic/Jurassic Boundary Vegetation Patterns and Climate
A compilation of palynological data from the Tr/J transition interval gives insight into the global distribution of palynofloral assemblages (Fig. 8, Appendix 2). Tr/J boundary sections from the Southern Hemisphere with abundant sporomorphs are scarce and often lack a high resolution. Therefore, we focus on the Northern Hemisphere. It is obvious that mixed assemblages consisting of pollen and spores in various abundances dominated the latest Rhaetian. In the earliest Hettangian, Classopollis-dominated assemblages spread across the interior of Pangaea, in areas affected by the basaltic volcanism of the Central Atlantic Magmatic Province (CAMP). Sections close to the Tethys Ocean show spore-dominated assemblages. During the Late Triassic, the Tethys, on the eastern edge of Pangaea, experienced a monsoonal climate because of the thermal contrast between the large continental plates along the equator and the sea (e.g., Parrish 1993; Satterley 1996; Sellwood and Valdes 2007). Evidence for Late Triassic monsoonal activity has been found, e.g., in the continental Chinle Formation, Colorado Plateau (Dubiel et al. 1991), from carbonate platforms in the southern Alps (Mutti and Weissert 1995), and from playa cycles in Germany (Reinhardt and Ricken 2000; Vollmer et al. 2008). The monsoonal activity influences precipitation patterns, and consequently floral distribution and development. Additionally, CAMP volcanism may have caused major climate change and substantial environmental disturbance, for example, from the increase of atmospheric CO2 concentration and associated warming (Marzoli et al. 1999; McElwain et al. 1999, 2009; Huynh and Poulsen 2005). The main pulses of the CAMP basalt correlate to the initial isotope shift and to the spore interval in St. Audrie's Bay (Deenen et al. 2010). We propose that the observed changes in palynofloral distribution and development are the indirect result of CAMP volcanism. The increase in greenhouse gases caused a warmer climate, the large Pangean landmass captured more heat, and an enhanced thermal contrast between the continent and the seas is the result. Consequently, the monsoon system got stronger and induced a drier interior and more intensive rainfall near the western margins of the Tethys Ocean.
Although some palynological studies report a change to a more humid climate (i.e., increase in fern spores) across the Tr/J boundary (Bonis et al. 2009; Götz et al. 2009), a change to more arid conditions on both sides of the North Atlantic rift during the earliest Jurassic is reflected by a palynofloral change to monotonous assemblages dominated by Classopollis (e.g., Fowell and Olsen 1993; Van Veen 1995; Whiteside et al. 2007; this study). Vakhrameev (1981) proposed the following classification based on the relative abundance of Classopollis: 1–10%, temperate climatic conditions; 20–50%, warm subtropical climate; and >60%, arid climate. Tr/J boundary sporomorph records studied by Hubbard and Boulter (2000) show pronounced and rapid climatic fluctuations culminating in a dramatic and protracted cooling event near the stage boundary. This cooling event is ambiguous because some botanical affinities were misinterpreted (e.g., C. meyeriana in the relatively cool group) and a lower stratigraphic resolution of these previous studies. It should be noted that some species of the Cheirolepidiaceae group (e.g., Classopollis torosus) may be indicative of coastal habitats (e.g., Batten 1974; Watson 1988; Abbink 1998). We propose that Classopollis meyeriana/Cheirolepidiaceae dominance indicates a drier climate in the earliest Hettangian and that the plants were capable of withstanding enhanced seasonality. “In general, one can conclude that that Cheirolepidiaceae were drought resistant, thermophilous shrubs and trees that required at least a subtropical climate” (Abbink 1998: p. 26). Further evidence of a dry Pangaean interior lies in the change from Late Triassic fluvial-lacustrine facies to early Jurassic eolian dune and interdune facies in southwestern United States (e.g., Tanner and Lucas 2007).
On the other hand, increased rainfall along the Tethys margin explains the shift to the Hettangian spore-dominated assemblages in Hochalplgraben. Wetter climate is also reflected in the lithology. Reefs were numerous in the latest Triassic Tethys (e.g., Stanley 2003; Seuβ et al. 2005; Kiessling et al. 2007). These need clear waters, and therefore runoff was probably negligible. The change to a wetter climate coincides with breakdown of carbonate production (reef extinction) (Hallam 2002; Hautmann 2004; Seuβ et al. 2005) and an increasing amount of silty marls, indicating enhanced runoff. Also the spore shift in the Tatra Mountains, Slovakia, is interpreted as displaying a sudden increase in humidity, most probably caused by the volcanic activity of the CAMP (Ruckwied and Götz 2009). Our data are in agreement with such a scenario. A recent study (Van de Schootbrugge et al. 2009) from the Germanic Basin showed a severe and wide ranging vegetation shift. Acid rain and noxious gases produced by CAMP volcanism may have devastated large regions of gymnosperm forests, which were subsequently replaced by pioneer assemblages consisting of ferns and lycophytes. However, their interpretation is not straightforward because the sections contain unconformities and the pollen record is likely biased by marked facies changes (Kürschner and Herngreen 2010 and references therein). The microfloral changes could be also the result of the climate changes as discussed above. It should be noted, however, that the Rhaetian–Liassic macrofloral record from the same region does not show any signs of a major mass extinction among plants (Kelber 2005; van Konijnenburg-van Cittert personal communication 2009). The lack of a major floral turnover in the macrofossil record supports the idea that in this case the palynological record may be biased by facies changes.
Higher-latitude records show differences in relative abundances and/or composition between the latest Rhaetian and earliest Hettangian but dominance by spores or Classopollis does not occur (Appendix 2). This is probably related to the higher latitude where a warm temperate climate prevailed (e.g., biome map fig. 5.14 in Willis and McElwain 2002) and the impact of the monsoon system was less severe. Continental Tr/J boundary deposits from the boreal realm (the northwest of the West Siberian Lowland) show sporomorph assemblages consisting of a variety of spores and pollen types (Rovnina 1972). There are no marked assemblage changes across the boundary. A remarkable difference compared to other Tr/J records is the high abundance of monosulcate pollen produced by Ginkgoaceae, Bennettitales, and Peltaspermales (Rovnina 1972), implying relatively “cooler” and humid conditions in the highest latitudes during the Tr/J transition interval. This is confirmed by the very low abundance of Classopollis.
The different distribution of palynological assemblages across the Tr/J boundary follows the climate changes inferred by Pangaean climate models. Mesozoic paleoclimate reconstructions generated on a general circulation model by Sellwood and Valdes (2007) show that the world was predominantly warm and that rainfall often focused over the oceans, leaving major desert expanses in the continental areas. There was no ice present at the high paleolatitudes (Frakes et al. 1992; Satterley 1996; Hallam and Wignall 1999). Huynh and Poulsen (2005) conducted a series of sensitivity experiments with a Late Triassic numerical coupled ocean/atmosphere climate model to predict extreme environmental conditions with 2–8 × pre-industrial CO2 levels. They found that on land, increasing CO2 caused extreme heating, intense seasonal fluctuations of surface temperatures, an increase in the number and severity of hot days and days without precipitation, and an exponential rise in the land surface area experiencing heat and aridity. Sensitivity experiments by Kutzbach and Gallimore (1989) also indicated a strong seasonality in low and middle latitudes, with seasonally intense precipitation and aridity. All of the experiments simulate dry conditions in the tropics (except along the east and west coasts), seasonally dry conditions in the midlatitude continental interiors, and year-round moist conditions only in middle or high latitudes (Kutzbach and Gallimore 1989). These model results are in agreement with the observed change to monotonous Cheirolepidiaceae vegetation in the continental interior of Pangaea and the fern-dominated assemblages near the Tethys Ocean.
Instead of a major and globally consistent palynofloral extinction event, the Tr/J boundary is characterized by climate-induced quantitative changes in the sporomorph assemblages that vary regionally in magnitude and composition. We propose that the changes in palynofloral distribution and the non-uniform development of pollen records are the result of climate changes related to CAMP volcanism. The increase in greenhouse gases caused a warmer climate and an enhanced thermal contrast between the continent and the seas. Consequently, the monsoon system became stronger and induced a drier interior and more intensive rainfall near the margins of the Tethys Ocean. The latest Rhaetian was dominated by mixed sporomorph assemblages consisting of pollen and spores in various abundances. Increased rainfall along the western Tethys margin explains the shift to the diverse Hettangian spore-dominated assemblages in Hochalplgraben. In contrast, the interior of Pangaea became drier and warmer. This can be seen, for example, in the St. Audrie's Bay section, where low-diversity Classopollis meyeriana assemblages dominated the record. Our results are in agreement with the Pangaean climate conditions inferred by climate model studies—a world that was predominantly warm with seasonal intense aridity and precipitation.
We acknowledge funding from the “High Potential” stimulation program of the board of Utrecht University. We thank J. van Tongeren and N. Welters for laboratory assistance. S. Hesselbo is thanked for providing us with samples from St. Audrie's Bay. M. Ruhl, M. Deenen, L. Krystyn, A. Von Hillebrandt, W. Krijgsman and M. Hounslow are thanked for their collaboration in the field and useful discussions. We gratefully acknowledge H. Visscher and H. Van Konijnenburg-Van Cittert for their comments on an earlier version of the manuscript. Constructive comments by D. Royer substantially improved the manuscript. This is publication no. 20110901 of the Netherlands Research School of Sedimentary Geology.
- Accepted 17 August 2011.