Ecological continuity and transformation after the Permo-Triassic mass extinction in northeastern Panthalassa

1. Introduction

The Permo-Triassic mass extinction (PTME) (251 Mya), the largest loss of biodiversity during the Phanerozoic Aeon, broke a long-dominant Paleozoic pattern of clade incumbency, resulting in a permanently altered marine realm [1]. While exact causes for the PTME are often debated, studies have indicated that this was the result of several variable and recurrent environmental stresses, such as high anoxia, hypercapnia, hydrogen sulfide and acidification—in addition to elevated temperatures and continental weathering—all of which coincided with ongoing climate warming and the eruption of the Siberian Traps [2–4]. The PTME precipitated the reorganization of Permian marine and terrestrial paleocommunities, leading to the emergence of new community types and biotic relationships [1,5]. Taxonomic and ecological dominance of invertebrate marine faunas shifted from the Paleozoic evolutionary fauna (PEF) (dominated by rhynchonelliform brachiopods, crinoids, rugose and tabulate corals, etc.) to the modern evolutionary fauna (MEF) (e.g. bivalves, gastropods, echinoids) [6]. While initially discussed in terms of taxonomic diversity and dominance, later studies have found that the MEF had greater functional diversity, ecological and morphological complexity, and an increased number of megaguilds, indicating a transformation to more active and diverse functional traits [7–9].

A central question of the transition from the PEF to the MEF is whether newly originated Triassic taxa performed new ecosystem functions, replaced functional roles vacated by victims of the PTME, and/or overlapped with surviving taxa because of shared biotic interactions or resource utilization. Here, we examine the functional diversity and niche space occupation of taxa from middle Permian (Guadalupian) and Early Triassic western North American marine paleocommunities to determine what ecological differences existed between taxa that became extinct during the PTME, those that survived the PTME and newly evolved Early Triassic taxa. Data were compiled from the published literature for the middle Permian (Wordian) Word Formation paleocommunity from Hess Canyon and Old Word Ranch (Texas) localities (see electronic supplementary material, extended data for sources) [10]. This paleocommunity was selected as a baseline for the PEF because it is temporally close to the Permian–Triassic boundary, is taxonomically rich with a well-preserved fauna and would not have been affected by the end-Guadalupian extinction event or the PTME [11]. Triassic data were compiled from field collecting and the published literature for three localities of the Spathian Virgin Limestone Member (Moenkopi Formation): Lost Cabin Springs in southern Nevada, and Hurricane Cliffs and Beaver Dam Mountains in southwestern Utah [10,12–15]. The Virgin Limestone Member has been interpreted to represent a low-diversity paleocommunity in the earliest phase of recovery from the PTME, commensurate with observations of Spathian recovery elsewhere [14–17], and as such was selected as a baseline for the emerging MEF.

While it has been previously hypothesized that stromatolites and microbialites present in the Virgin Limestone Member were formed during periodic anomalous conditions such as enhanced anoxia and alkalinity [18–20], a recent geochemical study indicated that anoxia was most likely not prevalent or sustained in the area [21], although this does not exclude the possibility of short-lived or recurrent deleterious conditions. Regardless, both the middle Permian and Early Triassic paleocommunities are parautochthonous and share similar depositional environments (mixed carbonate–siliciclastic sequences deposited on a shallow ramp during transgressive episodes), making them highly comparable [22–24]. Using an ordinated ecospace based on multiple functional traits of paleocommunity members, we examine whether there was indeed a shift from Paleozoic-style functional traits (e.g. passive, epifaunal, etc.) to more Modern-style traits (e.g. active, infaunal, etc.) in the aftermath of the PTME, as has been previously suggested.

2. Background and methods

Prior work proposed that new incumbents were established in the aftermath of the PTME as new taxa in previously unoccupied ecospace, as they are hypothesized to have evolved novel adaptations and niches [1,8]. Whether this transition between the PEF and MEF was protracted or abrupt, however, has been the subject of much debate. A number of studies suggest a protracted transition, as faunal components of both the PEF and MEF appear to have performed important functions during the Early and Middle Triassic [25,26]. It is important to note, however, that prior studies also focused heavily on taxonomic diversity patterns and global datasets [6,27,28], with less attention being paid to the functional differences between survivors and newly evolved taxa at local paleocommunity scales. Furthermore, taxonomic and functional diversity patterns are often decoupled after mass extinctions, highlighting the importance of examining both biodiversity components [29,30].

To examine the PEF–MEF transition, we applied a multi-trait functional framework, created by Novack-Gottshall [31], to each paleocommunity, aggregating genera into functional groups if they possessed identical traits across 10 functional categories, including body size, mobility, substrate relationship, etc. (electronic supplementary material, table S4) [31]. Taxonomic and ecological data were taken from the primary literature, Paleobiology Database (www.paleobiodb.org/) and field collections for the Early Triassic paleocommunity (see electronic supplementary material and extended data). Taxon abundance data were available for the Triassic paleocommunity owing to field collection, but not for the Permian Texan paleocommunity as the published literature used is mostly taxonomic-based, and hence did not include a count of individuals. Functional space occupation of each paleocommunity was estimated as the convex hull volumes of ordinated taxa based on principal coordinates analyses (PCOA) of inter-trait Gower dissimilarity matrices. Multivariate analysis of variance (MANOVA) and Hotelling's T² test were used to test the significance of the distributions of genera and individuals in multivariate space.

In addition, three components of functional diversity (richness, evenness and divergence) were calculated for each paleocommunity (see electronic supplementary material for details) [32]. Functional richness is the amount of niche space established by a paleocommunity, and is quantified alternatively as the number of functional groups present (F) or the PCOA hypervolume of occupied trait space (FRic). Functional evenness (FEve) is the evenness of species' or individuals' distributions in trait space, and functional divergence (FDiv) is the extent to which abundance distributions in niche space maximize divergence of functional characters within the paleocommunity. These measures are descriptive of the relationship between paleocommunity composition, ecosystem function and species diversity, and have proven useful for examining a variety of pre- and post-disturbance communities [32–34]. Meaningful descriptions of functional ecospaces, however, require integrated interpretations of these multiple descriptive metrics, which capture locations, coverage and distributions of taxa in those spaces.

3. Results and discussion

The Guadalupian paleocommunity was taxonomically diverse, with 125 genera including bivalves, brachiopods, ammonites, gastropods, corals, bryozoans, trilobites and echinoderms, whereas the Triassic had less than half that diversity, comprising 52 genera of bivalves, brachiopods, gastropods, ammonites and echinoderms. Interestingly, although the Guadalupian paleocommunity had more functional groups (F, Permian = 34; F, Triassic = 24), functional richness was greater in the Triassic paleocommunity (FRic, Permian = 0.78, Triassic = 0.95). By contrast, both communities had high functional divergence, measured as the mean distance of taxa from the centre of their distribution in ecospace (FDiv, Permian = 0.80, Triassic = 0.80). Therefore, whereas Guadalupian taxa performed a greater number of functions, Triassic taxa were more divergent in the functions that they performed. Both communities, however, were uneven, being dominated by two functional groups composed entirely of sessile, attached, suspension feeding genera (FEve, Guadalupian = 0.26, Triassic = 0.29).

Changes to functional structure caused by the PTME, and associated with the rise of the MEF, were examined by comparing three subsets of the Permian and Triassic taxa: (i) genera that became extinct during the PTME and whose functional roles were left vacant in the Early Triassic, (ii) PEF holdovers that survived into the Early Triassic, and (iii) Early Triassic taxa that originated after the extinction. The distribution of genera in occupied functional space differed significantly among the groups (MANOVA, Wilk's λ = 0.8410, p = 0.0091; figure 1), but this is attributable to a difference between extinct Permian taxa and holdovers only (Hotelling's T² = 20.638, p = 0.0011). This shows that the PTME was strongly selective, with attached epifaunal suspension feeding taxa having higher frequencies of extinction; this agrees with prior work suggesting that passive, non-motile taxa suffered higher extinction rates than those of active motile taxa [1]. Other examples of functional roles left vacant after the PTME, and unoccupied by new Early Triassic taxa, include those previously performed by deposit feeding bivalves and a trilobite, epibiont brachiopods and bivalves, small-sized carnivorous ammonites and an epifaunal grazing polyplacophoran. The ecospace distribution of newly originated Triassic genera, however, did not differ significantly from that vacated by the PTME extinctions (Hotelling's T² = 3.440, p = 0.5261) nor the surviving holdover taxa (Hotelling's T² = 5.826, p = 0.2369) because newly originated taxa both filled vacated functional roles and overlapped with those taxa that survived the PTME.

• Download figure
• Open in new tab
• Download PowerPoint

Despite the functional overlap of newly originated Triassic and Permian holdover taxa, the distributions of individuals, or abundance distributions, in trait space differed significantly between the two groups (Hotelling's T² = 1791.79, p < 0.0001, figure 2). Holdover taxa were numerically dominated by sessile, attached, suspension feeders (figure 2a). By contrast, newly originated Triassic taxa were dominated by mobile, semi-infaunal taxa (figure 2b). Thus, while the ecospace distributions of taxa do not differ between the two groups, new Triassic taxa were clearly dominated by those with more active MEF-style traits. This may indicate the beginning of the ecological transition and transformation of marine ecosystems to the characteristic MEF, a transition previously suggested to have been the result of collapse and passive replacement of the PEF by the MEF fauna [35]. Furthermore, functional evenness and divergence were calculated for newly originated taxa only (FEve = 0.24; FDiv = 0.80), revealing that these taxa were distributed unevenly, being clustered at the margins of represented ecospace. Thus, there was a high degree of niche differentiation in the Early Triassic, with coexisting species using resources in different ways, potentially allowing for more efficient resource use during recovery from the PTME [36].

• Download figure
• Open in new tab
• Download PowerPoint

The significance of species abundance in paleoecological studies continues to be debated [37,38], but many studies support abundance as an accurate indicator of ecological dominance [39–41]. We found that in terms of abundance, newly originated Triassic individuals were predominantly those with more active traits (i.e. mobile, semi-infaunal), potentially indicating the first step in the transition towards MEF-dominated ecosystems. Interestingly, Peters [42] reported similar results between the Cambrian and Ordovician, marking the transition between the Cambrian evolutionary fauna (CEF) and the PEF. There, trilobite-dominated Cambrian communities were replaced by brachiopod-dominated Ordovician communities [42]. While many more localities and paleoevironmental types need to be examined to determine if our findings are consistent with other PTME communities, the shift in functional emphasis we find across the PEF to MEF boundary may indicate the beginnings of changing metabolic requirements, biomass and morphology associated with the PTME and the rise of the MEF.

4. Summary

Newly evolved Early Triassic taxa of eastern Panthalassa refilled some functional roles vacated by the extinction of Permian taxa during the PTME. This is in broad agreement with prior work suggesting that taxa likely originate quickly and invade newly vacated ecospace during intervals of reorganization after mass extinctions [43]. However, newly originated Triassic taxa also overlapped functionally with PTME survivors, indicating that they shared traits and likely competed with surviving taxa. Surprisingly, newly originated Triassic taxa performed few novel ecosystem functions beyond those of middle Permian taxa. This is significant, as it is often assumed that taxa originating in the aftermath of mass extinctions expand into novel ecospace, establishing new stable faunal relationships early on as populations begin to interact in new ways [1].

PTME survivors and the newly originated MEF taxa were concentrated in different areas of functional space, with survivors dominated numerically by sessile, attached individuals and the MEF dominated by mobile, semi-infaunal individuals. Furthermore, the high functional divergence of the Triassic paleocommunity implies that, despite the overlap with Guadalupian taxa, Triassic individuals were at the extreme margins of the ecospace, potentially setting the stage for an expansion into novel ecospace via increasingly diverse functional traits [44]. Thus, although there was significant functional continuity between the PEF and MEF in the Early Triassic of western North America, functional emphasis shifted, potentially indicating the first step towards a lasting ecological and macroevolutionary transformation. Additional studies are needed to establish if the regional results reported here are typical of ecological recoveries after mass extinctions, or if each extinction imparts singular characteristics to its aftermath.

Data accessibility

Additional background, methods and taxa lists (with trait matrices) are available in the electronic supplementary material and in the Dryad Digital Repository: http://dx.doi.org/10.5061/dryad.bq25514 [10].

Authors' contributions

A.A.D., P.D.R. and M.L.F. conceived the study. A.A.D. and M.L.F conducted the fieldwork. A.A.D determined and assigned traits to genera. A.A.D. and P.D.R. conducted the statistical analysis. A.A.D. drafted the manuscript. All authors discussed the results, edited the manuscript, approved the final version and agree to be held accountable for the content.

Competing interests

The authors have no competing interests.

Funding

This work was supported by the Geological Society of America (GSA), The Explorers Club Exploration Fund, WI Geological Society, and the UW-Milwaukee Geosciences Department.