Benthic Foraminiferal Community Changes Across the Miocene Climatic Optimum Identified by Shebi Analysis (She Analysis for Biozone Identification), Calvert Cliffs, Maryland, USA

Extensive climate change occurred during the Middle Miocene. The Miocene Climatic Optimum (MCO; 17–14.8 Ma) is a potential analogue for a future Earth under the effects of anthropogenic climate change due to similar ocean configurations along with decreased area of global ice sheets, weaker equatorial trade winds, and a similar pCO₂ level to those estimated for 2050 (Tierney et al., 2020; Holbourn et al., 2022; Robinson et al., 2022; Hess et al., 2023). This climatic event is the warmest time of the Neogene with global mean annual bottom water temperatures (based on Mg/Ca ratios from benthic foraminifera) ∼4–8°C higher than preindustrial and with estimates of global sea level 35 m to 70 m higher than present (Miller et al., 2020). The degassing and eruption of the Colombia River Basalts, beginning around 17 Ma, would have put large amounts of CO₂ into the atmosphere and is the likely trigger for warming across the MCO (Kasbohm et al., 2018). During the subsequent Middle Miocene Climatic Transition (MMCT; 14.8–13.8 Ma), a sea-level fall to pre-MCO levels likely occurred with an expansion of the East Antarctic Ice Sheet as global temperatures decreased (Miller et al., 2020; Holbourn et al., 2022). A drawdown of CO₂ due to increased productivity and subsequent burial of organic carbon within the oceans across the Middle Miocene is potentially the cause of the MMCT (Compton et al., 1993; Flower & Kennett, 1993; Mallinson & Compton, 1997; Holbourn et al., 2022).

The Calvert Cliffs, MD, comprise a well-studied section of Miocene deposits composed of the Calvert, Choptank, and Saint Marys formations, which are known for well-preserved, rich marine fauna and well-preserved stratigraphic sequences (Figs. 1, 2; Kidwell et al., 2015). Strata within the Calvert and Choptank formations were deposited during the Miocene Climatic Optimum and the subsequent Middle Miocene Climate Transition (Fig. 2; Vogt & Parrish, 2012). Most paleontological studies of the Calvert Cliffs have focused on macrofossils from the rich shell beds and a bone bed, with fewer studies focused on the microfossils (Ward, 1992; Fuller & Godfrey, 2007; Vissaggi & Godfery, 2010; Kidwell et al., 2015; Carnevale & Godfrey, 2018; Culver et al., 2021). Previous studies on benthic foraminifera from Calvert Cliffs have focused on paleoenvironmental reconstructions of the strata, paleoecology, biostratigraphy, and community dynamics (Dorsey, 1948; Gernant, 1970; Buzas & Gibson, 1990; Culver et al., 2021), but none have used SHE analysis to investigate foraminiferal communities.

SHE analysis differentiates biofacies (communities) based on changes in species richness (the number of species in an assemblage; S), diversity (information function; H), and evenness (E), rather than on changes only in the relative abundance of species. The basis of SHE analysis is the decomposition equation [H = lnS+lnE]. Individuals are added together in a stepwise successive way over samples, while H, lnS and lnE are calculated at each step creating a linear trend. Breaks in these trends are what determines new biofacies (Hayek & Buzas, 1997; Buzas & Hayek, 1998).

This study used the foraminiferal assemblages from the Calvert Cliffs spanning the Calvert and lower Choptank formations to test the following hypotheses: 1) whether SHE analysis can differentiate different beds of the Calvert Cliffs based on benthic foraminiferal assemblages; 2) whether SHE analysis can differentiate strata deposited within the MCO and MMCT. By doing so, this study will document how benthic foraminiferal communities change across sea-level cycles and across an episode of climate change with similar pCO₂ levels to those projected for 2050 (Tierney et al., 2020).

During the Miocene, the Salisbury Embayment of the Atlantic Coastal Plain was covered by a shallow sea (Ward & Andrews, 2008; Zimmt et al., 2022). The major source of sediments was from the northwest creating thick deposits near central Maryland. These approximately 70 m thick deposits form the Chesapeake Group, which is composed of the Calvert Formation, the Choptank Formation, and the St. Marys Formation (Shattuck, 1902, 1904; Ward & Andrews, 2008; Powars et al., 2015). These formations were split into twenty-four informal stratigraphic units based on lithology and faunal content known as “Shattuck zones” (Shattuck, 1902), which will be referred to herein as “beds” following Kidwell et al. (2015).

Based upon macrofossil and microfossil assemblages, the beds of the Calvert Cliffs record a shallowing upward series of open marine, shelf environments within the Calvert and Choptank formations to fluvial deposits in the St. Marys Formation (Kidwell et al., 2015). Kidwell (1997) interpreted these beds within a sequence-stratigraphic context recognizing incised valley fills (IVF), transgressive system tracts (TST), and highstand system tracts (HST) that are bounded by unconformities (Fig. 2).

This study focuses on beds 4–9, which are considered a single unit following Ward & Andrews (2008), through 18: these beds span the MCO and MMCT (Vogt & Parrish, 2012; Zimmt et al., 2022). The Plum Point Member of the Calvert Formation comprises beds 4–9 through 16A. Beds 4–9 are incised valley fill deposits that were deposited in an inner shelf to shoreface environment (Kidwell, 1997). The overlying bed 10 is the first of four densely fossiliferous shell beds in the Calvert Cliffs, the others being beds 14, 17, and 19 (Kidwell et al., 2015). This bed has been split into the main body of bed 10, which is distinguishable by the dense mollusk assemblages, and the less fossiliferous and siltier bed 10 wedge (Fig. 2; Zimmt et al., 2022). Age estimates for the main body of bed 10 are 16.79–15.73 Ma (from the Willows locality), and the overlying wedge has been dated at 15.49–15.11 Ma (from the Plum Point locality) based on strontium isotope stratigraphy (Fig. 1C; Zimmt et al., 2022). Overlying bed 11 is a greenish blue sandy clay that was deposited in an inner shelf to outer shelf environment. Both beds 10 and 11 are considered to represent TSTs (Fig. 2; Kidwell, 1997; Kidwell et al., 2015).

Beds 12 and 13 both represent HSTs (Kidwell, 1997). Bed 12 is a thin, laterally extensive, bone-rich, brownish, sandy clay (Shattuck, 1904) deposited in an inner shelf to middle shelf environment and, according to Kidwell et al. (2015), represents the maximum water depth of the Calvert Cliffs strata. However, Culver et al. (2021) found this bed did not contain many planktonic foraminifera and was sandier than underlying bed 11 and overlying bed 13. Culver et al. (2021) concluded that bed 12 was most likely deposited in less than 50 m of water and represents the beginning of the shallowing after the maximum flooding surface between beds 11 and 12. The overlying bed 13 is a blueish, sandy clay deposited in an inner shelf environment (Fig. 2; Shattuck, 1904; Kidwell, 1988).

Bed 14 is a densely fossiliferous, sandy clay, deposited in a shoreface to inner shelf environment (Shattuck, 1904; Kidwell, 1984; Culver et al., 2021). This bed represents a TST that was deposited during the onset of the MMCT (Kidwell, 1984; Vogt & Parrish, 2012; Holbourn et al., 2015; Kidwell et al., 2015; Zimmt et al., 2022). The overlying bed 15 is yellow sand that was deposited in an inner shelf environment (Shattuck, 1904; Kidwell et al., 2015; Culver et al., 2021). Overlying bed 15 is bed 16A, which is the uppermost bed of the Calvert Formation (Fig. 2). This bed is a Turritella-rich, interbedded, silty sand and clay that, according to Kidwell et al. (2015), was deposited in an intertidal to low-energy subtidal environment and, along with bed 15, represents a HST (Fig. 2).

Bed 16B is the basal unit of the Choptank Formation and is known as the Governor Run sand (Fig 2; Kidwell et al., 2015). This bed is an IVF represented by yellow sand with mussels and sand dollars and was deposited in an intertidal environment (Fig. 2, according to Kidwell (1984, 1988, 1997)). However, based on benthic foraminiferal assemblages, Gernant (1970) argued this unit was deposited in the coldest waters of the Choptank at depths of 45 to 60 m. A more recent foraminiferal study suggests that shallow inner shelf deposits infilled the incised valley (Culver et al., 2021).

The overlying bed 17 is the Drumcliff Member of the Choptank Formation and is a TST (Fig. 2; Kidwell, 1997). This bed is recognized as a densely fossiliferous fine sand that was deposited in an inner shelf to shoreface environment at about 8–25 m water depth (Gernant, 1970; Kidwell, 1984, 1997; Kidwell et al., 2015). Overlying bed 17 is bed 18, which is the St. Leonard Member of the Choptank Formation. It is a blueish, green clay representing a HST (Fig. 2; Shattuck, 1904; Kidwell, 1997; Kidwell et al., 2015) and was interpreted by Kidwell (1984) to represent an intertidal to shoreface environment. Buzas & Gibson (1990) found that bed 18 was distinguishable by the high relative abundance of Bulimina elongata and Bolivina paula, two species generally considered to indicate decreased oxygen, high nutrient shelf environments (Katrosh & Snyder, 1982). Culver et al. (2021) recorded bed 18 as a sandy mud with a foraminiferal assemblage heavily dominated by Bolivna paula, in accordance with the findings of Buzas & Gibson (1990).

The stratigraphy of each section was logged according to the methods of Farrell et al. (2013). This study uses the previously published benthic foraminiferal data from Culver et al. (2021). Samples were taken from beds 4–9 through 18 at decimeter to meter scale (Fig. 3). The first few centimeters of sediment were removed from the outcrop to ensure weathered sediment was not collected. Sixty-three samples, three from each of the 21 sampling levels, were taken from outcrops at six different locations (Figs. 1C, 3). The 63 samples were divided into three laterally equivalent composite vertical sections of 21 samples each (Fig. 3). The distance between the three samples at each level varied from 1.8 m and ∼200 m. Strata were determined to be in the MCO or MMCT based on the ages of the beds from Vogt & Parrish (2012) and Zimmt et al. (2022), compared to the ages of the MCO and MMCT of Holbourn et al. (2015).

Each sample was washed over a nest of 63–710-μm sieves to remove the mud and coarse fraction. The remaining aliquot was then floated using sodium polytungstate to concentrate the foraminifera from the siliciclastic sediment (Munsterman & Kerstholt, 1996). The sink fraction was checked for foraminifera after floating to ensure that no specimens were lost. The remaining float fraction was then split using a micro-splitter until an adequate amount of sediment was put on a picking tray. A random numbers table was then used to identify squares to pick foraminifera until ∼300 foraminifera specimens were collected. Foraminifera were then identified using previously published literature on benthic fauna of the US Atlantic Coastal Plain (Cushman, 1918; Cushman & Cahill, 1933; Dorsey, 1948; McLean, 1956; Gibson, 1983; Snyder et al., 1988).

The previously published foraminiferal assemblage data of Culver et al. (2021) showed that the three composite sections were not statistically different in their species diversity metrics. The replicate samples for each section were combined by adding all specimen counts from each stratigraphic level to form one composite dataset of 21 samples (Appendices 1 and 2). Thus, each sample discussed below is a combination of three replicates. Preservation of foraminiferal tests was generally good. For example, specimens with thin, fragile tests, such as those from the genera Lagena and Fissurina, were identifiable to the species level. Specimens that could not be identified to the species level were poorly preserved (fragmented tests or tests diagenetically altered to the point that no distinguishable features could be identified); these were excluded from the dataset. A total of 84 species were included in the SHE analysis.

This equation shows there is a linear trend between these three measures. For example, if H is constant and lnS were to increase, then lnE (which is negative because evenness is always a value between zero and one) must decrease by the same amount. The other principle behind SHE analysis is that as the number of individuals in a population increases, the number of species should increase as well (Hayek & Buzas, 1997, 2010). This assumption is based on the log normal distribution of species where more specimens added to a community result in the documentation of more rare species (Buzas et al., 1982). SHE analysis adds the number of individuals to each subsequent sample and then calculates H, S, and E for each step. The natural logs of S and E are calculated then for each step and plotted with H against the natural log of the number of individuals. The variables lnS, lnE, and H will have a linear trend, and any breaks from this trend are considered to indicate a new biofacies or community (Buzas & Hayek, 1998). SHE analysis was carried out from the oldest strata to the youngest (Wilson, 2008).

Two communities (biofacies) were differentiated by SHE analysis in the 21 sets of combined samples (Fig. 4). The boundary between these two communities differentiates between strata deposited during the MCO (17–14.8 Ma) and the MMCT (14–13.8 Ma). The two uppermost samples of the section, the top of bed 17 and bed 18, were not able to be differentiated into communities because there were not enough samples to delineate trends in addition to the two communities that were recognized (Fig. 4).

Community 1 is composed of nine samples at the break of lnN = 9.02, or N = 8247 (Fig. 4, Table 1). These samples span beds 4–9 through bed 13, which encompasses all beds within the MCO (Fig. 5). These beds also represent a sequence of IVF, TST, and HST deposits. The community is composed of 8247 specimens and 58 species (Table 1). The cumulative H for this community is 2.56, and the cumulative evenness is 0.22 (Table 1). Five species dominate community 1, all of which have a mean relative abundance of >5%. Valvulineria floridana (29.2%) and Bulimina elongata (14.1%) are the two most dominant species followed by Cibicides americanus (8.5%), Bolivina paula (6.7%), and Hanzawaia concentrica (6.4%; Table 2).

Community 2 is composed of 10 samples at the break of lnN = 9.07, or N = 8693 and is encompassed within the MMCT at the Calvert Cliffs (Figs. 4, 5; Table 1). These samples are from beds 14 through 16B spanning a TST (bed 14), a HST (beds 15–16A), and an IVF (bed 16B). This community has a total of 8693 specimens and 74 species (Fig. 4; Table 1). The cumulative evenness for this community is 0.18 (Table 1). The dominant species in community 2 are Cibicides americanus (22.5%), Bolivina paula (15.0%), Valvulineria floridana (12.5%), Trochulina bassleri (8.4%), Buliminella elegantissima (7.2%), Lobatula lobatula (6.3%), and Bulimina elongata (5.3%; Table 2).

SHE analysis differentiated two communities. It did not identify individual beds based on benthic foraminiferal communities; neither community 1 nor community 2 were restricted to samples only from one bed. Instead, community 1 samples were from strata deposited during the MCO, and community 2 samples were from strata deposited during the MMCT. Thus, each climatic interval was characterized by different benthic foraminiferal communities across the strata deposited during these separate climatic events (Figs. 4, 5).

Community 1 characterizes all strata that were deposited during the MCO (Fig. 5). The co-occurrence of Valvulineria floridana, Cibicides americanus, and Hanzawaia concentrica in great abundance (Table 2) is diagnostic of an inner to middle shelf environment (Schnitker, 1971; Katrosh & Snyder, 1982; Snyder et al., 1988). Valvulineria floridana has a high relative abundance throughout community 1 (Fig. 5; Table 2). Peak relative abundance of this taxon occurs in beds 12 and the lower portion of bed 13, and then decreases upwards. The generally high relative abundance of this species throughout the studied section indicates a generally well oxygenated environment in an inner to middle shelf environment. The relative abundance of Cibicides americanus decreases up section throughout community 1, with the lowest abundance across beds 11–13 (Fig. 5). This species is most likely more indicative of shallow water environments, showing a deepening upward trend across the strata comprising community 1. Species belonging to Bulimina and Bolivina are generally characteristic of low-oxygen bottom water conditions (Poag et al., 1980; Sen Gupta & Machain-Castillo, 1993; Corliss & Chen, 1988). Bulimina elongata is generally found in high abundance in continental slope environments, but also can be found on the shelf in environments with high concentrations of nutrients (Schnitker, 1971; Poag, 1981; Katrosh & Snyder, 1982). In this study, the highest relative abundance of Bulimina elongata occurred in beds 11 and 13, the muddiest of the beds within the MCO (Fig. 5; Table 2). The high abundance of this species and the muddier sediments indicate these two beds were most likely deposited in the most nutrient-rich and deepest waters of the Calvert Cliffs section. The decrease in abundance of Valvulineria floridana coinciding with peaks in the relative abundance of Bulimina elongata is most likely due to different oxygen niches of these species (Fig. 5). In summary, the community composition indicated by SHE analysis is indicative of an inner to middle shelf environment that has an increased amount of nutrients in the system causing lower-oxygenated bottom waters (Fig. 5; see also Kidwell et al., 2015, and Culver et al., 2021). This differs from environmental interpretations presented by Kidwell et al. (2015) and Culver et al. (2021) because those studies interpreted individual beds, whereas SHE analysis grouped several beds together (Fig. 5).

No community boundary changes were detected by SHE analysis within strata deposited during the MCO. Foraminiferal communities at the onset of the MCO and throughout its peak temperatures were grouped together. This differs from the Paleocene Eocene Thermal Maximum (PETM) event that has been studied using SHE analysis. Hayek et al. (2019) developed and used a new Perturbation Detection Analysis, which incorporated a SHE analysis of 29 subsamples of benthic foraminiferal assemblage data from the onset of the PETM at ODP Site 690 to recognize benthic foraminiferal communities, before determining whether communities were in decline, growth, or stasis. SHE analysis recognized two communities, one immediately before the PETM carbon isotope excursion (CIE) and another during the CIE (Hayek et al., 2019). The PETM differs from the MCO in its more severe temperature excursion and its faster rate of onset (McInerney & Wing, 2011). The inclusion of all MCO samples in one community in the SHE analysis of this study could be a product of the longer time benthic foraminiferal communities had to adapt to environmental changes compared to the PETM.

Community 2 occurs in strata deposited during the MMCT (Fig. 5). Sediments of these strata are generally coarser than those of strata containing community 1, indicating a shallower depositional environment (Fig. 5). Community 2 has dominant species (>5%) in common with community 1, although in different rank order (Cibicides americanus, Bolivina paula, Valvulineria floridana, and Bulimina elongata), but the list of dominant taxa (>5%) also includes Trochulina bassleri, Buliminella elegantissima, and Lobatula lobatula (Table 2). Lobatula lobatula is found at depths less than 200 m along the modern Atlantic coast with distributions generally concentrated in shallower coastal to inner shelf environments (Culver & Buzas, 1980). Buliminella elegantissima can also be found from shallow coastal environments to the continental slope in modern assemblages, but it reaches high abundance in shallower shelf environments with higher nutrient input (Culver & Buzas, 1980; Katrosh & Snyder, 1982). The inclusion of Buliminella elegantissima in these facies with Bolivina paula and Bulimina elongata supports a continuation of bottom waters with lowered oxygen concentrations (Poag et al., 1980; Corliss & Chen, 1988; Sen Gupta & Machain-Castillo, 1993). The relative abundance of Bolivina paula increases throughout beds 15 and into 16B within community 2 (Fig. 5). These beds are much sandier than those of beds 11 and 13 within community 1 which have a greater abundance of Bulimina elongata, thus indicating higher nutrients in shallower water (Fig. 5). However, the epifaunal taxa, Cibicides americanus and Lobatula lobatula, are indicative of more normal oxygenated marine environments (Katrosh & Snyder, 1982; Corliss & Chen, 1988). The dominant taxon, Cibicides americanus, increases in relative abundance as the taxa Valvulineria floridana decreases in relative abundance at bed 15 (Fig. 5). The coarser sediments and decrease in the relative abundance of Valvulineria floridana, which has been described by Snyder & Katrosh (1982) as a species diagnostic of a middle shelf environment, indicates community 2 is shallower than community 1 but with similar elevated nutrient levels (see also Kidwell et al., 2015, and Culver et al., 2021; Fig. 5). As in community 1, SHE analysis did not differentiate individual beds within the interval that included beds 14 to 17, leading to more general paleoenvironmental interpretations than those of Kidwell et al. (2015) and Culver et al. (2021) (Fig. 5).

In summary, SHE analysis is shown by the analyses of this paper to be a useful tool in recognizing benthic foraminiferal community changes across intervals of global climate change. Community change in the Miocene strata at the Calvert Cliffs occurs between samples from beds 13 and 14 (∼25 cm apart) and is congruent with the temperature decrease and sea level fall associated with the MCO and MMCT boundary.

SHE analysis differentiated two communities within beds 4–9 through 18 from the Calvert Cliffs, MD. Community 1 was deposited during the MCO, and community 2 was deposited during the MMCT. Community 1 (beds 4–9 through 13) contained species that indicated an inner to middle shelf environment with increased nutrient input and possibly lower oxygen. Community 2 (beds 14 through 17) contains species that generally reflect a shallower inner shelf environment. These results show the efficacy of SHE analysis in recognizing community changes related to Neogene climatic events.

We would like to thank Peter Vogt, Stephen Godfrey, Kevin Foley, and Susan Kidwell for sharing their knowledge of the Calvert Cliffs. Also, thank you to Mimi Katz, Jim Browning, and Rob Poirier for their comments that helped improve this paper. SRS would like to acknowledge Dr. Martin A. Buzas for his guidance and enthusiasm when the research for this paper was in its infancy. This research is part of the U.S. Geological Survey Geologic Investigations of the Neogene project. Finally, thank you to the Cushman Foundation for two student research grants that funded SRS. The Appendices can be found linked to the online version of this article.