Plant megafossils from the Paleocene-Eocene Thermal Maximum (PETM) in the Bighorn Basin, north-central Wyoming, USA, document a dramatic shift in floral composition, whereas palynofloral change from the same sections has appeared to be more subtle. We investigated this discrepancy by quantifying pollen preservation and measuring the stable carbon isotope composition of specific pollen taxa. Pollen grains belonging to two common latest Paleocene taxa are poorly preserved in PETM samples, and their δ13Cpollen is similar during the latest Paleocene and PETM. In contrast, pollen grains of a thermophilic taxon that became more abundant during the PETM are pristine, and the δ13Cpollen of PETM specimens is ∼4‰ lower than that of latest Paleocene specimens. More broadly, pollen grains belonging to lineages currently centered in temperate climates are poorly preserved when found in PETM samples, whereas in the same samples, pollen belonging to lineages now centered in the tropics are well preserved. These differences in preservation and isotopic composition indicate extensive redeposition of older pollen grains during the PETM. Increased abundance of Cretaceous palynotaxa in PETM samples confirms erosion and redeposition, likely resulting from more episodic and intense precipitation. Exclusion of reworked palynotaxa from analyses reveals that, as in the megaflora, temperate taxa were absent during the PETM at the time when dry tropical taxa briefly appeared. Major climate changes like the PETM may commonly destabilize landscapes, increase reworking, and thus smooth patterns of change in microfloras, leading to underestimates of the rate and magnitude of floral response to past global change.

During the Paleocene-Eocene Thermal Maximum (PETM), ca. 56 Ma, a massive release of isotopically light carbon to the ocean-atmosphere system generated a negative carbon isotope (δ13C) excursion (CIE) of 4‰–5‰ and global warming of 4–8 °C (McInerney and Wing, 2011). Prior work on megafloras across the PETM in the Bighorn Basin, northcentral Wyoming, USA, showed major change in composition across the hyperthermal event (Wing et al., 2005; Wing and Currano, 2013). Nearly half of 91 common megafloral taxa had no occurrences in the CIE body, although they occurred both before and after the PETM in the study area (range-gap taxa). Nearly a quarter of the megaflora occurred only during the PETM (PETM-only taxa). Range-gap taxa typically have living relatives distributed in temperate climates, whereas living relatives of PETM-only taxa grow in dry tropical climates. Wing and Currano (2013) interpreted range gaps as local extirpations caused by hotter, drier climate and PETM-only distributions as geographic range expansions responding to the same climate change. Although Wing et al. (2005) found rare occurrences of tropical palynomorphs (e.g., Brosipollis, Bombacoideae) in the PETM, they did not find range-gap palynotaxa, even among plant families for which megafossil records showed range gaps (e.g., Cupressaceae).

The absence of range gaps in the palynoflora might reflect the greater sampling radius of typical palynomorph assemblages compared with highly local leaf assemblages (Jacobson and Bradshaw, 1981; Ferguson, 1985), implying that local populations of range-gap taxa did persist in the basin, but not close to environments of leaf deposition. Alternately, range gaps in the palynoflora might have been closed by Paleocene palynomorphs being eroded and redeposited during the PETM, a process that leaf fossils are too fragile to undergo (Ferguson, 1985).

We used the preservational state and stable carbon isotopic composition of range-through palynotaxa to test the hypothesis that they were reworked from Paleocene sediments. Paleocene pollen grains redeposited during the CIE body should reflect the stable carbon isotopic composition of the Paleocene atmosphere rather than the PETM atmosphere because pollen preserves the δ13C signature of atmospheric CO2 at the time of growth (Farquhar et al., 1989; Jahren, 2004). Such grains should also be less well preserved than autochthonous PETM pollen.

Geologic Setting

The Paleocene Fort Union Formation and Paleocene–Eocene Willwood Formation of north-central Wyoming are well exposed in the northwest-southeast–oriented Bighorn Basin (BHB). The PETM has been identified in many parts of the BHB as a 30–45-m-thick sequence of fluvial rocks and paleosols (Koch et al., 1992; Bowen et al., 2001; Rose et al., 2012; Baczynski et al., 2013; van der Meulen et al., 2020). We collected plant microfossils from five types of fluvial subenvironments characterized by lithology, internal structure, geometry, proximity to paleochannels, and size (Table S1 in the Supplemental Material1). The age of each sample was interpolated from its stratigraphic position relative to horizons of known age (e.g., magnetic reversals, onset of the CIE, end of the CIE body), assuming uniform long-term rates of sediment accumulation between dated levels. Previous work established that the CIE body represents ~100 k.y. (Westerhold et al., 2018; van der Meulen et al., 2020), had mean annual temperatures ~5 °C higher than the latest Paleocene (Fricke and Wing, 2004; Wing et al., 2005), and had a drier or more seasonally dry climate (Wing et al., 2005; Kraus and Riggins, 2007; Kraus et al., 2013; Baczynski et al., 2016). The CIE recovery is estimated to have been ~40 k.y. in duration and was characterized by decreasing temperature and increasing floodplain wetness (Wing et al., 2005; Adams et al., 2011; Baczynski et al., 2016; van der Meulen et al., 2020).


Palynomorphs were recovered from 40 samples (Fig. S1; Table S2) using a modified standard palynological process (see the Supplemental Material for additional details). Every fossil palynomorph was scored as corroded, degraded, and/or broken, following the terms of Cushing (1967) (Fig. S2). The classes are not mutually exclusive. Thermal maturity of grains was assessed using the palynomorph darkness index (PDI) of Goodhue and Clayton (2010). Palynomorphs affected by two or more forms of damage were considered deteriorated.

Carbon Isotope Analysis

For each individual carbon isotope measurement of pollen (δ13Cpollen), 30–40 grains of the selected pollen taxon were identified under light microscopy from the residue of a single sample. These grains were transferred by micropipette to a drop of ultrapure water that was combusted using a spooling-wire microcombustion device. The δ13Cpollen values of the resulting CO2 gas were measured using an isotope ratio mass spectrometer (IRMS). The methods are described more fully in the Supplemental Material and in Nelson (2012). Weight percent total organic carbon (%TOC) and carbon isotopic measurements of TOC (δ13CTOC) were measured using an elemental analyzer IRMS as described by Baczynski et al. (2013). The n-alkane carbon isotope values (δ13Cn-C29) were determined via gas chromatography IRMS using methods described in the Supplemental Material and by Baczynski et al. (2016, 2018).

Pollen Preservation

Samples from the CIE body had high proportions of deteriorated palynomorphs and dinoflagellates compared with those from before and after, though four samples from the last 150 k.y. of the Paleocene had intermediate levels (Fig. 1). Deteriorated grains in PETM samples included clearly reworked Mesozoic taxa (e.g., Aequitriradites ornatus, Classopollis classoides, Aquilapollenites quadrilobus, and marine dinoflagellates; Table S3) as well as those with Paleocene and younger ranges (e.g., Caryapollenites veripites, Momipites leffingwellii, and Ulmipollenites krempii). The proportion of dinoflagellates and deteriorated grains varied by depositional environment, but abandoned channel fills, the only environment distributed throughout the sequence, showed an increase during the CIE body (Fig. 1).

We examined the preservation of two temperate palynotaxa, the conifer Cupressacites hiatipites and the walnut relative Caryapollenites veripites, during the PETM, as these palynotaxa are relatively common in Paleocene rocks and therefore more likely to be reworked. More than 75% of C. hiatipites and C. veripites grains from the CIE body were deteriorated, whereas <5% from before and after were deteriorated (T value <0.02; Fig. 2; Table S8). In contrast, the palm Arecipites tenuiexinous, which likely preferred warmer climates, is pristinely preserved throughout the sequence.

Carbon Isotope Measurements

The carbon isotope composition of bulk organic matter (δ13CTOC) ranged from -30.1‰ to -23.9‰ through the sequence. The median values were -25.7‰ for the Paleocene, -28.2‰ for the CIE body, -27.0‰ for the CIE recovery, and -26.1‰ for the post-PETM Eocene (Fig. 1). During the CIE body, δ13CTOC values were found to be positively correlated with the proportion of deteriorated palynomorphs in samples (R2 = 0.76, P value < 0.0010; Fig. S3).

The carbon isotope composition of C29 n-alkanes (δ13Cn-C29) showed the expected 4‰–5‰ negative CIE associated with the PETM in addition to the sudden onset and rapid recovery, consistent with previous observations (Fig. 1; Table S6; Baczynski et al., 2013).

Each measurement of δ13Cpollen was made on 30–40 grains of a specific pollen type from a single rock sample. The error of individual measurements was calculated from the mass of carbon in the sample (Fig. 3; see the Supplemental Material). If we measured more than one group of pollen grains from a sample, we present their error-weighted mean value. The mean δ13Cpollen of late Paleocene C. hiatipites was -22.6‰ (N = 2), and the mean δ13Cpollen of C. veripites from the same sample was -26.1‰ (N = 2). Latest Paleocene A. tenuiexinous had a mean δ13Cpollen of -26.8‰. C. hiatipites from the CIE body had δ13Cpollen of -23.5‰ (N = 1), and C. veripites from the CIE body had δ13Cpollen of -26.6‰ (N = 1). A. tenuiexinous from the CIE body had a δ13Cpollen of -30.8‰ (N = 4; Fig. 3; Table S4).

The carbon isotope composition and preservational state of palynomorphs are consistent with the hypothesis that many organic microfossils were eroded from Paleocene (and older) rocks and redeposited during the CIE body in the Bighorn Basin. The δ13Cpollen of C. hiatipites and C. veripites shifted only fractionally and insignificantly negative from the latest Paleocene to the CIE body, consistent with most or all of the grains being products of photosynthesis from the Paleocene rather than PETM atmosphere (Fig. 3). Specimens of these taxa from the CIE body were also generally poorly preserved (Fig. 2), consistent with erosion from (slightly) older sediments and redeposition during the CIE body. In contrast, the δ13Cpollen of A. tenuiexinous shifted ~–4‰ from the latest Paleocene to the CIE body, consistent with photosynthesis from a CIE atmosphere ~4‰ more negative than the latest Paleocene atmosphere, which agrees with the magnitude of the CIE observed in n-alkanes and mammalian tooth enamel (Fig. 3; Koch et al., 1992; Secord et al., 2012; Baczynski et al., 2016). Specimens of A. tenuiexinous from the CIE body were also well preserved and commonly occurred in clusters, providing independent evidence they were autochthonous and not eroded, transported, and redeposited from older rocks (Farley, 1990).

We considered alternative hypotheses to explain the absence of a negative shift in the δ13C values of C. veripites and C. hiatipites. Each of these palynotaxa might represent more than one closely related biological species (see the Supplemental Material; pollen of related species may not be distinguishable). If the hypothetical species within C. veripites and C. hiatipites that lived during the PETM had ~4‰ less preference for 12C than the Paleocene species producing the same pollen types, this would offset the CIE in the atmosphere and result in similar δ13Cpollen values in the Paleocene and PETM, as we observed. Similarly, changes in photosynthesis and/or stomatal conductance, resulting from decreased water availability, could have reduced discrimination against 13C in plants growing during the CIE (Diefendorf et al., 2010). These alternative hypotheses, though, require species and/or physiologic changes in unrelated plant groups that exactly offset the difference between the pre-CIE and the CIE body carbon isotope values, while a third species, A. tenuiexinous, displayed a carbon isotopic change very similar to that observed in δ13CTOC and δ13Cn-C29. These alternative hypotheses also do not explain the poor preservational state of the former two taxa during the CIE. The reworking hypothesis is more parsimonious and therefore preferred here.

Other evidence for elevated reworking during the CIE body is seen in higher proportions of deteriorated palynomorphs and Cretaceous dinoflagellates. Taxonomic identifications (Table S3) of the reworked palynomorphs and dinoflagellates suggest the dominant sources for fine reworked organic matter were the easily eroded Cretaceous and Paleocene mud rocks around the Bighorn Basin margin. Both the source and increased amount of reworking are consistent with prior studies of the isotopic and maceral composition of dispersed organic matter in multiple sections (Bataille et al., 2013; Baczynski et al., 2013, 2016, 2019; Denis et al., 2021). We do not attribute the increase in the proportion of reworked palynomorphs to decreasing production of pollen and spores during the PETM because sedimentary concentrations of organic microfossils remained high, at least in some depositional environments (Table S5).

By evaluating the preservational state of palynological assemblages with light microscopy, we recognized nine palynotaxa that appeared to range through the PETM but were represented only by deteriorated, likely reworked, grains (Fig. 4). Removing these taxa from pollen sums revealed higher turnover in the PETM palynoflora than previously recognized and, specifically, that, as in the megaflora, there are range gaps for palynotaxa during the CIE body. These range-gap taxa represent 37% of the palynoflora, comprising a smaller proportion than in the megaflora, but far more than previously recognized. Significantly, the range-gap palynotaxa with known nearest living relatives belong to plant groups with high extant diversity in temperate climates (e.g., Aesculus, Alnus, Carya, Platanus). The patterns of turnover in the palynoflora are more congruent with the megaflora than previously recognized, and they bolster the hypothesis that many temperate plants were extirpated from the Bighorn Basin during the CIE body. New taxa from the CIE are illustrated and described by Korasidis et al. (2022).

Recently, the isotopic composition of individual planktic foraminifera tests at Ocean Drilling Program Site 865 in the mid-Pacific Ocean was used to demonstrate that turnover in faunal composition across the PETM was muted by mixing (Hupp et al., 2022). Although there is abundant evidence for increased flux of terrestrial organic matter in marine deposits along many continental margins during the PETM (Handley et al., 2008; Carmichael et al., 2017; Lyons et al., 2019), there has been less attention paid to the possibility that erosion and redeposition could have reduced the apparent rate and magnitude of biotic turnover in organic microfossils. Unrecognized reworking of pollen and spores may disguise the distinctiveness of PETM palynofloras in many sections, and, more broadly, similar effects during other periods of rapid environmental change may reduce the apparent magnitude and rapidity of floral response to major disruptions in Earth history.

1Supplemental Material. Expanded methods, taxonomic considerations, Tables S1–S8, and Figures S1–S3. Please visit to access the supplemental material, and contact with any questions.

Funding for this research was provided through a Smithsonian Institution Postdoctoral Fellowship Award and a Paleontological Society Norman Newell Grant to V.A. Korasidis. We are grateful to three anonymous reviewers for their constructive and helpful reviews on the manuscript. P. Bijl is also thanked for his assistance with dinoflagellate identification.

Gold Open Access: This paper is published under the terms of the CC-BY license.