Geological control on dinosaurs’ rise to dominance: Late Triassic ecosystem stress by relative sea level change

The Late Triassic was a time of significant biotic upheaval, with the origination of new groups such as dinosaurs, lizards, crocodiles and mammals, but also characterized by a prolonged period of extinctions, distinguishing it from other great mass extinction events, while the gradual rise of the dinosaurs during the late Carnian to Norian remain unexplained (Benton, Forth, & Langer, 2014; Brusatte, Nesbitt, et al., 2010; Sereno, 1999). This stepwise, important shift in terrestrial life occurred over a prolonged period with complex patterns of mass extinctions (Bernardi, Gianolla, Petti, Mietto, & Benton, 2018; Lucas & Tanner, 2018; Tanner, Lucas, & Chapman, 2004). Each extinction event was characterized by distinct turnovers in flora and fauna, but these events have all been attributed to different forcing factors, including the Carnian Pluvial Event (CPE), triggered by Wrangellian volcanism (Dal Corso et al., 2012). Putative extinction events through the Norian have been explained by bolide impacts (Clutson, Brown, & Tanner, 2018). Importantly though, other distinct extinction events during the Late Triassic (Brusatte, Nesbitt, et al., 2010) are not associated with any known external forcing factors. Because of this complexity, the very premise of a singular end-Triassic mass extinction event has been questioned (Hallam & Wignall, 1999; Lucas & Tanner, 2018; Rigo et al., 2020). Low-gradient delta plains are key habitats that have been instrumental in the evolution of life throughout Earth history by acting as shelters for surviving species following environmental crises and providing an arena for interaction between the marine and terrestrial realm (Greb, DiMichele, & Gastaldo, 2006). Lethally hot equatorial temperatures at the onset (Sun et al., 2012) of, and periodically through (Whiteside et al., 2015), the Triassic placed extra emphasis on these important deltaic refugia, but also meant that parts of the world normally not crucial in the evolution of life (Jablonski, Roy, & Valentine, 2006) became more important (Spalletti, Artabe, & Morel, 2003)—as demonstrated by prolific tetrapod faunas north of 30°N and south of 40°S during the Triassic (Lucas, 2018). At such northern latitudes, vast delta systems developed within the Received: 23 January 2020 | Revised: 17 April 2020 | Accepted: 28 May 2020 DOI: 10.1111/ter.12480


| INTRODUC TI ON
The Late Triassic was a time of significant biotic upheaval, with the origination of new groups such as dinosaurs, lizards, crocodiles and mammals, but also characterized by a prolonged period of extinctions, distinguishing it from other great mass extinction events, while the gradual rise of the dinosaurs during the late Carnian to Norian remain unexplained (Benton, Forth, & Langer, 2014;Brusatte, Nesbitt, et al., 2010;Sereno, 1999). This stepwise, important shift in terrestrial life occurred over a prolonged period with complex patterns of mass extinctions (Bernardi, Gianolla, Petti, Mietto, & Benton, 2018;Lucas & Tanner, 2018;Tanner, Lucas, & Chapman, 2004). Each extinction event was characterized by distinct turnovers in flora and fauna, but these events have all been attributed to different forcing factors, including the Carnian Pluvial Event (CPE), triggered by Wrangellian volcanism (Dal Corso et al., 2012). Putative extinction events through the Norian have been explained by bolide impacts (Clutson, Brown, & Tanner, 2018). Importantly though, other distinct extinction events during the Late Triassic (Brusatte, Nesbitt, et al., 2010) are not associated with any known external forcing factors. Because of this complexity, the very premise of a singular end-Triassic mass extinction event has been questioned (Hallam & Wignall, 1999;Lucas & Tanner, 2018;Rigo et al., 2020).
Low-gradient delta plains are key habitats that have been instrumental in the evolution of life throughout Earth history by acting as shelters for surviving species following environmental crises and providing an arena for interaction between the marine and terrestrial realm (Greb, DiMichele, & Gastaldo, 2006). Lethally hot equatorial temperatures at the onset (Sun et al., 2012) of, and periodically through (Whiteside et al., 2015), the Triassic placed extra emphasis on these important deltaic refugia, but also meant that parts of the world normally not crucial in the evolution of life (Jablonski, Roy, & Valentine, 2006) became more important (Spalletti, Artabe, & Morel, 2003)-as demonstrated by prolific tetrapod faunas north of 30°N and south of 40°S during the Triassic (Lucas, 2018 Triassic Boreal Ocean (TBO) and have been mapped across areas unmatched by any analogue in Earth's history (Klausen, Nyberg, & Helland-Hansen, 2019).
Terrestrial palynomorph records document the change in dominant plants in a given geographical area and are routinely used as a relative age proxy in addition to being considered indicators of climate and depositional environment. Extensively documented miospore assemblages from the western Barents Sea region record the first occurrence of several new taxa at the Carnian-Norian transition. These include Cingulizonates rhaeticus, Kyrtomisporis gracilis, K. laevigatus, K. speciosus, Limbosporites lundbladiae and Retitriletes austroclavatidites (Paterson & Mangerud, 2015. The botanical affinity of Kyrtomisporis is enigmatic but it is generally interpreted as a fern spore, whereas Cingulizonates, Limbosporites and Retitriletes are considered to represent lycopsid spores (Bonis & Kürschner, 2012) Collectively, these species reflect a parent plant flora, which grew in a warm temperate climatic belt along the northern Pangaean coast (Boucot, Chen, & Scotese, 2013;Sellwood & Valdes, 2006).
Palynological investigations (Hochuli & Vigran, 2010;Paterson & Mangerud, 2015;Paterson et al., 2016;Vigran, Mangerud, Mørk, Worsley, & Hochuli, 2014) have revealed that significant compositional differences exist between the composition Late Triassic palynofloras of the TBO and those of the 'classic' Germanic and Alpine areas (Table S1). This includes both differences in the relative abundances of various palynomorph groups (Vigran et al., 2014) and in the stratigraphical ranges of key species and events (Mueller, Hounslow, & Kürschner, 2016). The former has been related to palaeoclimatic control, whereas the latter has been tentatively linked  to the earlier origination of several plant genera in the TBO during the late Carnian-early Norian.
In this study, we show how the palynological record of the TBO compares to other basins around the world and explain the terrestrial turnover by significant shifts in major deltaic systems. This terrestrial turnover reinvigorates old ideas about extinctions in the Late Triassic and we hypothesize that cyclic turnovers in terrestrial environments could explain the punctuated rise to dominance of Dinosauria.

| RE SULTS
The first appearance of the spore taxa C. rhaeticus, K. gracilis, K. laevigatus, K. speciosus, L. lundbladiae and R. austroclavatidites in the western Barents Sea region is recorded in the uppermost parts of the De Geerdalen Formation and correlative Snadd Formation (Paterson & Mangerud, 2015. This interval is assigned to the late Carnian (Tuvalian) by magnetostratigraphy (Lord et al., 2014), which is calibrated to the global time-scale by an early Norian ammonite fauna from the overlying Flatsalen Formation (Korčinskaya, 1980). Elsewhere in the TBO, such as in the Sverdrup Basin of Arctic Canada, comparable palynomorph assemblages are recorded from lower Norian deposits (e.g. Fisher & Bujak, 1975;Suneby & Hills, 1988). Kyrtomisporis gracilis is endemic to the TBO.
While the other spore species listed are commonly represented in Late Triassic assemblages from other basins in the northern Pangaea, their first occurrence elsewhere is significantly delayed. For instance in the more arid continental interior regions, such as the UK and Germany, these taxa do not appear in the fossil record until the late Rhaetian (Morbey, 1975;Lund, 1979;Kürschner & Herngreen, 2010) ( Figure 1a), coinciding with an apparent humid phase (Ahlberg, Arndorff, & Guy-Ohlson, 2002;Götz, Ruckwied, & Barbacka, 2011;Hesselbo, McRoberts, & Pálfy, 2007). The origination of these species in the TBO region ( Figure 1b) therefore predates their earliest consistent occurrences within the Germanic Realm by up to 20 million years (Figure 2a). Although the distribution of these palynomorphs is facies-controlled to some degree, it seems probable that their parent plants first evolved in the TBO during the late Carnian, before spreading southwards into the interior of Pangaea in the Rhaetian.
The appearance of these species in the TBO coincides with a major floral turnover and paleoenvironmental change in the region.
At the maximum regressive stage, slightly prior to their appearance, the deltaic flora over this vast area was dominated by the 'tree like' fern Asterotheca meriani (Pott, 2014), as reflected by an acme of its spore Leschikisporis aduncus (Hochuli & Vigran, 2010;Paterson & Mangerud, 2015;Paterson et al., 2016. The abrupt regional extinction of this species in the western Barents Sea during the latest Carnian coincides with marine inundation of the delta, an increased abundance of palynomorphs derived from putative halophytic plants, brackish water algae, agglutinated foraminifera and, in the latter stages, marine microphytoplankton (Mangerud, Paterson, & Riding, 2019;Paterson et al., 2016). The recorded turnover implies significant environmental stress and loss of terrestrial habitats were potential drivers for the evolution of these forms, suggesting they were produced by opportunistic plant species that were able to colonize the new niches created by the widespread flooding of the delta.
Evidence for severe habitat loss is clearly seen in, and explained by, the geomorphological characteristics of the TBO: stratigraphic intervals constrained by palynological ranges comprise deltaic and shallow marine environments that can be traced from Northeast Russia to Svalbard and beyond ( Figure 1b). Delta plains within TBO covered more than 1.65 × 10 6 km 2 by conservative estimates (Klausen et al., 2019) at stages of maximum regression-roughly equal to 1% of total land areas in the modern world ( Figure 1b).
These widespread delta plains facilitated vast low-gradient coastal habitats, but inherently also made them highly susceptible to marine inundation during periods with increased eustatic sea level or altered sediment supply.
Gradients are usually very low on the tops of large-scale deltas (Paola et al., 2011), and although these landforms re- In the prevailing Greenhouse setting of the Triassic (Retallack, 2013;Sellwood & Valdes, 2006), RSL was primarily controlled by tectonics, climate and sediment supply changing over longer time-scales than modern glacially driven RSL. These controls likely had a global impact on coastal regions around the globe, as reflected in the eustatic signal in basins from different parts of Pangea (Haq, Hardenbol, & Vail, 1988). Flooding of the low-gradient TBO F I G U R E 1 Global Late Triassic paleogeography and present day setting of study area. a Paleogeographic setting and climate belts of Triassic Pangaea (Boucot et al., 2013). Global record of palynological assemblages indicates a diachronous distribution of flora. The Triassic Boreal Ocean (orange box) are studied in the subsurface Barents Sea and outcrops on Svalbard b, located at higher latitudes than in the Triassic. Wells and seismic surveys used in the present study are indicated, and the hatched pattern indicates the minimum extent of the TBO delta system at maximum regression (Klausen et al., 2019)  delta plain was therefore less frequent than modern, but, because of its size, much more extensive. Cyclic or episodic changes in sediment supply add to the potential of major RSL change in TBO.
The main difference was the fully bipedal posture of the former that was more advantageous in the terrestrial habitats that terrestrial species were confined to during sea level highstand. As RSL later dropped, vast deltaic habitats would be open for recolonization by species formerly confined to more marginal hinterlands.
The rise of Dinosauria as the dominant clade during the Late Triassic has been attributed to a corresponding decrease in the importance of crurotarsans, but explained by several factors: competitive advantage (Brusatte, Benton, Ruta, & Lloyd, 2008); changing environmental conditions (Brusatte et al., 2008); and altered atmospheric CO 2 (Schaller, Wright, & Kent, 2015;Whiteside et al., 2015).
Competitive advantage as a causal mechanism has been refuted (Benton et al., 2014;Brusatte et al., 2008), and there are no direct links between pCO 2 and extinctions at the Carnian-Norian boundary ( Figure 4). The most convincing indications of a link between abiotic stress and terrestrial life is the recent study linking Late Triassic F I G U R E 4 Global climatic and evolutionary events plotted against observations from TBO. Relative sea level observed in wells from the Barents Sea (e.g. Figure 3d) with a complete record of the Triassic interval (Figure 3), regional and global events affecting sedimentation pattern and the evolution of life (Dal Corso et al., 2012;Whiteside et al., 2015), pCO 2 Royer, Berner, Montañez, Tabor, & Beerling, 2004), and terrestrial extinction events (Paterson & Mangerud, 2015). This record shows that although large volcanic events had important effects on the evolution of life, and climatic crises likely affected diversity (Bernardi et al., 2018), there are significant turnovers in the terrestrial realm at stages with no apparent change in pCO 2 , volcanism, climatic events or any large bolide impact (Clutson et al., 2018). TBO RSL changes however tie directly to each major crurotarsan extinction, as indicated by red triangles. Note that extinctions of dinosaurs do not exhibit the same correlation to RSL as do crurotarsan species. CAMP = Central Atlantic Magmatic Province (Peters, 2005) In.

J.
Het. diversification of dinosaurs to the CPE (Bernardi et al., 2018), but this study fails to explain the numerous important extinctions and diversification stages after this climatic crisis (Lucas & Tanner, 2018).
At the Carnian-Norian boundary, this study records a pronounced terrestrial turnover in TBO with no known corresponding volcanic (Dal Corso et al., 2012) or bolide events (Clutson et al., 2018).
The sole agent for extinction at this stage is ecological stress caused by marine inundation. Importantly, flooding seems to closely correspond to the last occurrence of several crurotarsan lineages also at other stages in the Triassic (Figure 4).Were the crurotarsans not able to migrate at pace with the receding shoreline and keeping up with the changing conditions? Although TBO transgressions likely span much shorter periods than regressions, implying they were relatively catastrophic events even within million-year cycles, decrease in habitat area must be regarded as more important than the pace at which it happen. As in the Anthropocene, land-loss could not have been a trivial matter for species depending on these habitats.
However, the effect of sea-level change on Late Triassic extinctions has largely been considered to affect marine organisms and the discussion has been focused on the extinction at the Triassic-Jurassic boundary (Hallam & Wignall, 1999;Lucas & Tanner, 2018;Tanner et al., 2004). An additional major difference between our observations and previous studies is that we correlate last occurrence of terrestrial species to marine transgressions rather than the regressions proposed by previous studies.
Based on the terrestrial turnover in TBO at a time with no other external forcing factor, we propose that RSL played a much more important role for turnovers in the terrestrial realm than previously accepted. Correlation furthermore shows that cyclic RSL changes could also explain the disappearance of other crurotarsan lineages in the Triassic. Unlike other agents of extinction that seem to be unique to a given period and potentially coincidental (Bernardi et al., 2018;Clutson et al., 2018;Whiteside et al., 2015), RSL change re-occur at periods and with a timing that directly corresponds to important turnovers in the terrestrial realm ( . 4). It is also a theory that can explain the gradual and stepwise disappearance of crurotarsans without a significant extinction event, but rather link the last occurrence of dinosaurs' main competitors to turnovers in the terrestrial habitat they were specialized for and dominated.

| CON CLUS IONS
Investigated palynological species have significantly earlier first occurrences within TBO, later dispersing to the Germanic and Tethyan realm-contrary to normal nucleation-dispersion trends. The largest delta system in Earth's history developed in the TBO at this time and its geomorphological character made it prone to widespread transgressions seen at multiple stratigraphic intervals, causing discrete turnovers in flora at maximum flooding. Inundations imply significant habitat loss and ecosystem stress impacting species adept to these areas, such as the crurotarsans. Distinct stepwise decrease in ecosystem importance suffered by the crurotarsans correlate directly to discrete transgressions worldwide, and although some of these extinctions can also be correlated to other forcing factors, all of them can be linked to a distinct flooding event. The Carnian-Norian interval studied herein however is not associated with any other external factor but flooding, showing that environmental stress by relatively rapid marine inundation played a crucial role in the evolution and turnover of terrestrial habitats. We record cyclic devastation of vast delta plains in the Boreal Ocean, which represented large and important habitats for species such as crurotarsans that dominated the Triassic terrestrial realm, and suggest that this ecological stress facilitated the gradual and implicit rise to ecosystem dominance by Dinosauria.

ACK N OWLED G EM ENTS
We thank the Norwegian Research Council for funding through the ISBAR project, grant number 267689. Thoughtful comments and suggestions from two anonymous reviewers and editor Max Coleman improved the quality of the manuscript and is much appreciated.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available in the DISKOS database courtesy of the Norwegian Petroleum Directorate.
Palynological data used to support the findings in the study are derived from literature review, listed in the supplementary material.