Biofacies Analysis of Zanclean Sediments in Virginia: Unraveling the Past Through Benthic Foraminifera

Abstract

Early Pliocene sedimentary deposits are exposed at the surface along the James and York Rivers, across southeastern Virginia. The Zanclean age (5.33–3.60 Ma) Sunken Meadow Member of the Yorktown Formation records a relatively large-scale marine transgression in the Salisbury Embayment. A total of 15 samples were collected from an outcrop near Spring Grove, VA, for grain-size analysis and to document benthic and planktic foraminiferal assemblages. The sediments are generally moderately well-sorted, shelly fine sands. A total of 48 benthic taxa were recorded from the Sunken Meadow Member, though only 14 taxa occurred in proportions high enough to be included in the Q-mode cluster analysis (>3% of the total assemblage). Low numbers of planktic foraminifera indicate relatively shallow water deposition. Biofacies analysis shows three distinct biofacies groups in the Sunken Meadow Member and the benthic foraminiferal community shifts throughout the unit are indicative of changes in nutrient availability, surface productivity, and bottom water oxygenation. The results indicate a middle to outer neritic depositional environment similar to modern conditions found south of Cape Hatteras, NC.

1. Introduction

Climate models project rates and levels of warming unprecedented in Earth’s recent history by the end of this century [1]. Global mean surface temperatures (GMSTs) have increased over the last several decades at rates unseen on century to millennial timescales [2]. Increases in global temperature and atmospheric CO₂ are having widespread impacts on both terrestrial and marine environments [1,2], increasing the need for better understanding of how these systems respond to warming. Deep-time paleoclimate records preserve regional to global scale boundary conditions similar to future scenario projections and provide perspective into potential environmental changes that could occur by the end of the 21st century [2,3,4].

Pliocene (5.33–2.59 Ma) warm periods, particularly the mid-Piacenzian Warm Period (MPWP, 3.264–3.025 Ma), have been the basis of global climate reconstructions for roughly three decades due to atmospheric and physical surface conditions that are nearly analogous to end-of-century projections [3,4,5]. Atmospheric CO₂ concentrations during these warm intervals are estimated to be between 350 and 450 ppm [3,6,7], comparable to current CO₂ levels and end-of-century projections but significantly elevated compared to pre-industrial (1750–1800 AD) concentrations of ~280 ppm [8]. Current (2023 AD) global concentrations of atmospheric CO₂ have reached ~419 ppm, a higher level than any time in the last 2 million years [2,9]. Similarities in extant species, ocean circulation, land mass distribution, and continental positioning are additional factors in making the Pliocene a popular analogue to near-future conditions [3,10,11].

A succession of Neogene transgressions are well exposed along the mid-Atlantic Coastal Plain (MACP), particularly along the James River, in southeastern Virginia, USA (Figure 1). The Yorktown Formation records three phases of sea-level rise across the MACP during the Pliocene [12,13]. These sediments were deposited at a time when the global average sea level and temperatures were ~25 m and 3 °C higher than during the pre-industrial [3,4,10,11,14], and atmospheric CO₂ concentrations were similar to present (~400 ppm; [6,7]).

The Zanclean age (5.33–3.60 Ma) Sunken Meadow Member is the basal unit of the Yorktown Formation, recording a large marine transgression in southeastern Virginia and northeastern North Carolina [12]. Age estimates of the Sunken Meadow Member place the unit between 4.8 and 3.8 Ma [14,15,16,17]. This transgression occupied a large portion of the Virginia coastal plain, though it is smaller and less extensive than the transgressive phase that resulted in the overlying Rushmere–Morgarts Beach members [12,18]. The MPWP, which occurs prior to the transition from Marine Isotope Stage (MIS) M2 to M1, is recorded in Rushmere and Morgarts Beach members [4].

Due to the dynamic nature of sea-level changes, continental shelf deposits are temporally incomplete compared to deep-sea settings, often recording only warm (sea-level highstand) climate conditions [12]. The Yorktown Formation has a long history of study [4], and these sediments give insight into global sea-level change, Pliocene global ice volume, and how shallow marine ecosystems respond to warmer climate states in the western North Atlantic. Benthic foraminifera have been described from the York-James Peninsula by Cushman [19] and McLean [20], though these studies were performed prior to the stratigraphic revision by Ward and Blackwelder [10]. Previous regional study of the Sunken Meadow Member has focused on the molluscan and ostracod assemblages [10,21,22]. The only extensive work on the benthic foraminiferal assemblage of the Sunken Meadow Member was conducted on exposures at Lee Creek Mine, in Aurora, NC [23,24].

2. Materials and Methods

Fifteen samples were collected for grain-size and benthic foraminiferal assemblage analysis from a section of the Sunken Meadow Member exposed at the Pipsico Scout Reservation in Spring Grove, VA, in December 2021 (Figure 1). The section is located along the James River, where Miocene to Pliocene sediments are accessible in bluffs of variable thickness.

2.1. Benthic Foraminifera

The exposed section was scraped to remove weathered material, and samples were collected every 15 cm from the accessible portion of the unit. Sample spacing was determined in the field, based on sediment fabric and outcrop condition. Lateral replicate samples were not obtained due to current erosion of the bluffs. Roughly 200 to 600 g of sediment per sample was collected and dried in a gravity convection oven at ≤50 °C overnight. After drying, samples were weighed and agitated in a dilute (5 gL⁻¹ water) solution of sodium hexametaphosphate for approximately 1 h. Samples were washed over a 63 μm mesh sieve until all muds were removed and specimens were clean. After washing, samples were dried at ≤50 °C, and approximately 300 benthic foraminiferal specimens were obtained from dried sample material after splitting samples with a microsplitter. The entire > 63 μm sediment aliquot was then picked for benthic and planktic foraminifera using a gridded picking tray. Benthic species were then identified by comparison with other published works [19,20,21,23,24,25,26,27,28,29]; planktic species were not identified due to a paucity of specimens and small test size.

Percentages of Textulariina, Miliolina, and Rotaliina were calculated for each sample. The relative abundance of planktic foraminifera was calculated using the standard formula in Equation (1):

$$\%\mathrm{P}=({\displaystyle \frac{\mathrm{P}}{\mathrm{P}+\mathrm{B}}})\times 100$$

(1)

where P is equal to the number of planktic foraminifera specimens picked and B is equal to the number of benthic foraminifera specimens picked.

Relative abundance of benthic species was calculated from foraminiferal census data for all 15 samples. Q-mode cluster analysis [30] was used to distinguish groups within relative abundance data. Q-mode clustering was chosen over R-mode to compare each sample population instead of comparing species-to-species relationships, preserving the time-stratigraphic aspect of the sample data. Relative abundance data were transformed using Equation (2):

$$2arc\mathrm{s}\mathrm{i}\mathrm{n}\sqrt{p_i}$$

(2)

where p_i is equal to the fraction of the ith species [31] prior to analysis. Only species represented by 3% or more of the total assemblage in one or more samples were included in the cluster analysis dataset. All samples were analyzed using Ward’s linkage and Euclidean distances in PAST4 [32]. Species richness (S), Shannon diversity (H), and evenness (E) were calculated for all samples using PAST4 [32] and Equations (3) and (4):

$$\mathrm{H}=-\sum _i{\displaystyle \frac{n_i}{n}}\ln {\displaystyle \frac{n_i}{n}}$$

(3)

$$E=e^{\mathrm{H}}∕\mathrm{S}$$

(4)

where n_i is the number of individuals in a species, n is the total number of individuals in the community, and S is the number of species [32,33]. The diversity index Fisher’s alpha is defined using Equation (5):

$$\mathrm{S}=a\times \ln (1+n/a)$$

(5)

where S is the number of taxa, n is the number of individuals, and a is the Fisher’s alpha [32,33].

2.2. Grain-Size Analysis

Thirty-gram aliquots were obtained for grain-size analysis from bulk sample material. Aliquots were washed using the same procedure outlined above. Once dry, samples were sieved over half-phi intervals from −2 phi to 4 phi (ϕ) using a mechanical sieve shaker. Weight percent of the >63 μm size fraction was calculated for each sample and analyzed using the GRADISTAT worksheet [34,35]. Weight percentages of mud (<4 ϕ) were estimated using the difference between the bulk and washed sample weights.

2.3. Alkenones

Corresponding paleotemperature data derived from alkenone paleothermometry are available from this field site [36]. Alkenone paleothermometry uses the $U{\displaystyle \frac{K^{\prime }}{37}}$ saturation index and the temperature dependence of double and triple carbon bonds in alkenones produced by haptophyte algae [4,37,38,39]. These data are calibrated using the Müller coretop calibration [40], with a calibration uncertainty of ±1.38 °C [39]. Details of the techniques used to generate data and other relevant information about these samples can be found in the associated metadata file [36].

3. Results

The Sunken Meadow Member is unconformably bound between the underlying Miocene Cobham Bay Member of the Eastover Formation and the overlying Late Pliocene Rushmere Member of the Yorktown Formation [10]. Regionally, across southeastern Virginia, this unit is generally a greenish-gray, shelly sand averaging ~3 m in thickness with a slight dip [10]. Visual distinction between the Sunken Meadow and Rushmere members in outcrop is difficult due to similarities in lithology and sedimentary structures. While not proper lithostratigraphic procedure, often the presence of the molluscan species Chesapecten jeffersonius, a marker species for Zone 1 Yorktown of Mansfield [41], is used to distinguish the two units [4,10]. At the field site for this study, the Sunken Meadow is a ~2 m thick section of light tan sand, lacking visible sedimentary structures; large mollusks are common throughout the unit.

3.1. Sedimentological Data

The Sunken Meadow Member at Spring Grove is composed of well- to moderately well-sorted, slightly gravelly fine sands (

Table 1. Table of grain-size analysis values. Samples are in stratigraphic order.

Sample Name	Mean (Mz)	Sorting (σI)	Mode	Mean	Sorting	Skewness	Kurtosis
21MR42	2.19	0.73	2.24	Fine sand	Moderately sorted	Very coarse skewed	Extremely leptokurtic
21MR41	2.22	0.69	2.24	Fine sand	Moderately well sorted	Coarse skewed	Very leptokurtic
21MR40	2.24	0.80	2.24	Fine sand	Moderately sorted	Very coarse skewed	Very leptokurtic
21MR39	2.29	0.59	2.24	Fine sand	Moderately well sorted	Coarse skewed	Very leptokurtic
21MR38	2.35	0.65	2.24	Fine sand	Moderately well sorted	Coarse skewed	Very leptokurtic
21MR37	2.31	0.82	2.24	Fine sand	Moderately sorted	Very coarse skewed	Very leptokurtic
21MR36	2.36	0.74	2.24	Fine sand	Moderately sorted	Very coarse skewed	Very leptokurtic
21MR35	2.45	0.60	2.24	Fine sand	Moderately well sorted	Coarse skewed	Very leptokurtic
21MR34	2.44	0.72	2.24	Fine sand	Moderately sorted	Very coarse skewed	Very leptokurtic
21MR33	2.47	0.69	2.74	Fine sand	Moderately well sorted	Very coarse skewed	Very leptokurtic
21MR32	2.43	0.67	2.24	Fine sand	Moderately well sorted	Coarse skewed	Very leptokurtic
21MR31	2.50	0.56	2.74	Fine sand	Moderately well sorted	Very coarse skewed	Very leptokurtic
21MR30	2.52	0.54	2.74	Fine sand	Moderately well sorted	Very coarse skewed	Very leptokurtic
21MR29	2.55	0.45	2.74	Fine sand	Well sorted	Coarse skewed	Leptokurtic
21MR28	2.57	0.55	2.74	Fine sand	Moderately well sorted	Very coarse skewed	Very leptokurtic

; Figure 2). The sands are predominately sub-rounded quartz grains with minor amounts of micas and glauconite. The gravel-sized fraction is comprised of large bioclastic grains (molluscan shell pieces); several shell beds occur in the section (Figure 3).

All samples are unimodal with a mode ranging from 2.24 ϕ to 2.74 ϕ (Table 1). The percent sand values plot between ~60 and 80% (Figure 2); gravel percentages range between ~10 and 40% (Figure 2). The mean grain-size of the >63 μm size fraction ranges from ~2.5 phi (ϕ) at the bottom of the section to 2.1 ϕ at the top (Table 1, Figure 2). The overall distribution of grains is coarsely to very coarsely skewed and generally very leptokurtic (Table 1). The values for sorting (σ₁) plot between ~0.4 and 0.8σ, with higher values generally occurring near the top of the unit (Table 1, Figure 2).

3.2. Foraminiferal Assemblages

The percentage of planktic foraminifera, in relation to all planktic and benthic foraminifera, does not exceed ~3% in the Sunken Meadow Member at Spring Grove (

Table 2. Values for foraminiferal assemblage characteristics by sample. Samples are in stratigraphic order. N, number of specimens picked; S, number of species; Bf/g, calculated number of benthic foraminifera per 1 g of sediment; Pf/g, planktic foraminifera per 1 g of sediment; %P, percent of planktic foraminifera; %t, percent of Textulariids; %m, percent of Miliolids; %r; percent of Rotaliids; α, Fisher’s alpha; H, Shannon’s diversity index; E, evenness.

Sample Name	N	S	Bf/g	Pf/g	%P	%t	%m	%r	α	H	E
21MR42	317	23	923.76	14.57	1.55	2.84	0.00	97.16	5.7	2.07	0.35
21MR41	731	27	817.57	13.49	1.62	3.44	0.00	96.56	5.52	2.34	0.39
21MR40	381	30	525.82	7.68	1.44	1.95	0.00	98.05	7.45	2.42	0.38
21MR39	417	31	443.67	9.58	2.11	2.88	0.00	97.12	7.74	2.45	0.38
21MR38	542	37	316.15	5.83	1.81	0.92	0.00	99.08	8.99	2.90	0.49
21MR37	332	30	170.47	3.49	2.06	2.71	0.00	97.29	8	2.88	0.59
21MR36	273	29	211.21	2.37	2.50	0.73	0.00	99.27	8.21	2.56	0.45
21MR35	370	26	605.87	8.60	0.54	1.35	0.00	98.65	6.38	2.11	0.32
21MR34	293	23	705.94	4.62	2.01	1.71	0.00	98.29	5.85	2.19	0.39
21MR33	288	29	877.11	9.61	2.02	0.77	0.00	99.23	7.25	2.15	0.30
21MR32	408	25	867.04	12.75	1.21	1.47	0.00	98.53	5.88	2.07	0.32
21MR31	278	22	715.76	7.56	3.14	2.52	0.00	97.48	5.61	1.98	0.33
21MR30	400	22	630.15	3.94	0.74	1.25	0.00	98.75	5.01	1.89	0.30
21MR29	693	28	1666.86	12.66	0.86	1.15	0.00	98.85	5.86	2.01	0.27
21MR28	310	27	202.57	7.12	1.90	4.19	0.00	95.81	7.11	2.55	0.48

; Figure 3b). Calculated numbers of planktic foraminifera per 1 g of sediment range from ~2 to 14 across the unit, with the highest numbers occurring at the top of the unit (Table 2). The overall benthic foraminiferal assemblage is dominated by Rotaliids, ranging from ~95% to 100% (Table 2); relative proportions of test type are plotted in Figure 3b. Agglutinated foraminifera do not exceed ~4% in any sample and occur in the highest numbers near the bottom of the unit (Table 2; Figure 3); miliolids are absent from the assemblage at Spring Grove (Table 2; Figure 4). Fisher’s alpha (α) values range from ~5 to 9 (Table 2, Figure 4). Values for H and E range from ~1.9–2.9 and ~0.27–0.59, respectively (Table 2, Figure 4). Species diversity is lower near the bottom of the section, and highest towards the middle of the section (Table 2). Of the 48 total taxa that comprise the foraminiferal assemblage, only 14 taxa occur in proportions greater than 3% of any singular sample and are included in the Q-mode cluster analysis (Section Cluster Results).

Cluster Results

The dendrogram resulting from Q-mode cluster analysis of the normalized relative abundance of benthic foraminiferal data is shown in Figure 3a. The 15 samples clustered into three groups (Figure 3a,

Table 3. Mean percent abundance and range for all taxa in biofacies groups determined by Q-mode cluster analysis of benthic foraminiferal relative abundance data.

Group 1	Mean		Group 2	Mean		Group 3	Mean
9 Samples, 41 Taxa	%	Range	2 Samples, 39 Taxa	%	Range	4 Samples, 37 Taxa	%	Range
Globocassidulinasubglobosa	28.84	19.41–36.86	Rosalina bassleri	17.27	15.36–19.19	Rosalina bassleri	27.29	22.54–39.12
Rosalina bassleri	21.48	13.87–30.88	Globocassidulinasubglobosa	16	13.55–18.45	Globocassidulinasubglobosa	18.89	11.04–23.03
Epistominelladanvillensis	21.05	13.55–29.00	Buliminellaelegantissima	8.28	5.72–10.84	Epistominelladanvillensis	17.81	14.39–20.77
Rosalina floridana	4.27	3.03–7.10	Buliminaelongata	7.51	3.87–11.14	Hanzawaiaconcentrica	6.20	3.79–8.63
Buliminellaelegantissima	2.79	0.75–6.59	Buccellafrigida	6.23	5.54–6.93	Buliminellaelegantissima	3.86	1.92–6.62
Cibicideslobatulus	2.59	0.52–6.77	Hanzawaiaconcentrica	5.28	3.92–6.64	Rosalina floridana	2.95	0.32–5.52
Textularia spp.	1.68	0.73–4.19	Epistominelladanvillensis	5.26	4.80–5.72	Textularia spp.	2.78	1.95–3.44
Buccellafrigida	1.65	0.49–4.40	Cibicideslobatulus	3.67	1.81–5.54	Cibicideslobatulus	2.44	0.00–5.28
Cibicidesamericanus	1.56	0.00–5.48	Elphidium spp.	3.33	1.85–4.82	Trifarinaillingi	2.01	0.00–7.79
Bolivinapaula	1.49	0.00–2.56	Bolivinabrevior	2.55	2.40–2.71	Bolivinapaula	1.97	1.22–1.92
Hanzawaiaconcentrica	1.45	0.36–2.73	Cibicidesamericanus	2.04	1.51–2.58	Buccellafrigida	1.96	0.96–1.89
Buliminaelongata	1.27	0.00–4.03	Textularia spp.	1.82	0.92–2.71	Cibicidesfletcheri	1.48	0.00–2.34
Bolivinabrevior	1.03	0.26–2.20	Cibicidesfletcheri	1.71	1.20–2.21	Elphidium spp.	1.32	0.49–1.65
Elphidium spp.	1	0.00–3.30	Fursenkoinafusiformis	1.7	1.29–2.11	Pseudoparrellapontoni	1.01	0.00–1.89
Cassidulinalaevigata	0.68	0.00–1.29	Bolivinapaula	1.62	1.20–2.03	Bolivinabrevior	0.95	0.49–1.10
Pseudononionpizarrensis	0.57	0.00–3.87	Rosalina floridana	1.62	1.20–2.03	Cassidulinalaevigata	0.77	0.49–0.72
Parafissurinabidens	0.56	0.00–1.80	Bolivina spp.	1.52	0.92–2.11	Buliminaelongata	0.73	0.24–0.83
Buccelladepressa	0.53	0.00–1.61	Buccelladepressa	1.4	1.29–1.51	Parafissurinabidens	0.60	0.00–0.96
Bolivinafloridana	0.51	0.00–1.72	Cassidulinalaevigata	1.34	1.20–1.48	Buccellainusitata	0.53	0.24–0.83
Bolivina spp.	0.49	0.00–1.71	Buliminellacurta	1.27	0.74–1.81	Fursenkoinafusiformis	0.50	0.00–0.96
Buccellainusitata	0.49	0.00–1.10	Discorbiscavernata	1.22	0.60–1.85	Trifarinaoccidentalis	0.47	0.14–1.20
Cibicidesfletcheri	0.46	0.00–1.89	Parafissurinabidens	1.04	0.60–1.48	Bolivina spp.	0.46	0.00–0.96
Trifarinaillingi	0.46	0.00–1.61	Trifarinaillingi	0.74	0.00–1.48	Buccelladepressa	0.44	0.24–0.63
Pseudoparrellapontoni	0.43	0.00–1.80	Lagena sulcata	0.64	0.37–0.90	Buliminellacurta	0.41	0.24–0.63
Fursenkoinafusiformis	0.43	0.00–2.20	Pseudoparrellapontoni	0.6	0.00–1.20	Pseudononionpizarrensis	0.35	0.00–0.69
Buliminellacurta	0.35	0.00–0.72	Trifarinaoccidentalis	0.6	0.00–1.20	Cibicidesamericanus	0.34	0.00–0.49
Valvulineriafloridana	0.34	0.00–1.29	Reussoolinalaevis	0.58	0.55–0.60	Nonionellamiocenica	0.26	0.00–0.73
Rosalina spp.	0.26	0.00–2.32	Pseudononionpizarrensis	0.52	0.30–0.74	Valvulineriafloridana	0.26	0.00–0.48
Reussoolinalaevis	0.24	0.00–0.50	Nonionellamiocenica	0.46	0.00–0.92	Bolivinafloridana	0.18	0.00–0.72
Trifarinaoccidentalis	0.21	0.00–0.68	Bolivinafloridana	0.43	0.30–0.55	Reussoolinalaevis	0.15	0.00–0.32
Nonionellamiocenica	0.17	0.00–0.74	Virgulina spp.	0.39	0.18–0.60	Bolivinaplicata	0.12	0.00–0.49
Bolivinaplicata	0.16	0.00–0.77	Buliminainflata	0.28	0.00–0.55	Discorbiscavernata	0.12	0.00–0.24
Patellinaadvena	0.09	0.00–0.49	Lagena substriata	0.28	0.00–0.55	Guttulinaaustriaca	0.12	0.00–0.48
Guttulinaaustriaca	0.08	0.00–0.37	Buccellainusitata	0.24	0.18–0.30	Lagena sulcata	0.12	0.00–0.48
Cancrissagra	0.07	0.00–0.36	Bolivinacalvertensis	0.18	0.00–0.37	Dentalina spp.	0.06	0.00–0.24
Virgulina spp.	0.06	0.00–0.29	Globulinainaequalis	0.09	0.00–0.18	Discorbis spp.	0.06	0.00–0.24
Astrononiongallowayi	0.05	0.00–0.49	Lagena hexagona	0.09	0.00–0.18	Virgulina spp.	0.03	0.00–0.14
Discorbis spp.	0.03	0.00–0.27	Nodosariacatesbyi	0.09	0.00–0.18
Nodosariacatesbyi	0.03	0.00–0.27	Strigialifususrutilus	0.09	0.00–0.18
Strigialifususrutilus	0.03	0.00–0.26
Globulinainaequalis	0.03	0.00–0.25

) at a Euclidean distance of ~1.5. Group 1 is composed of the bottom nine samples and 41 taxa (Figure 3a, Table 3), with Globocassidulina subglobosa (28.84%), Rosalina bassleri (21.48%), and Epistominella danvillensis (21.05%) as the most abundant taxa. Group 2 is composed of two samples and 39 taxa (Table 3), with 4 taxa in relatively high abundance but none dominating the assemblage: Rosalina bassleri (17.2%), G. subglobosa (16.00%), Buliminella elegantissima (8.28%), and Bulimina elongata (7.51%). Group 3 is composed of the four uppermost samples and 37 taxa (Figure 3a, Table 3), dominated by Rosalina bassleri (27.29%) with G. subglobosa (18.89%), E. danvillensis (17.81%), and Hanzawaia concentrica (6.20%) in relatively high abundance.

3.3. Alkenone Sea Surface Temperatures

Sea surface temperatures (SST) derived from the $U_{37}^{K^{\prime }}$ saturation index range between 25.97 and 27.76 °C (Figure 4) across the section. The bottom eight samples (21MR11–18) range from 27.6 °C to 27.74 °C; four samples above that are ~27 °C. There is an abrupt 1 °C drop in temperature, from 27 °C to ~26 °C, across three samples, before returning to 27.74 °C at the top of the unit.

4. Discussion

The Sunken Meadow Member has been interpreted as being deposited in an open neritic, mild temperate environment during an early Pliocene transgression [10,42,43,44,45]. This unit is often massive in outcrop, lacking in visible sedimentary structures and bioturbation, with relatively limited exposures at the surface in southeastern Virginia, and can be difficult to identify in subsurface cores [4]. Global shifts in climate during the Pliocene are recognized as widespread, diachronous marine deposits across the MACP. Large molluscan shells are abundant throughout the unit, often occurring in ~0.5 m thick layers. The age of the Sunken Meadow Member within the Zanclean is debated and has gone largely unconstrained due to a lack of biostratigraphically important microfossils [15,16,17,43]. Refining the age of the unit is beyond the scope of this work, which aims to build on previous descriptions of the Sunken Meadow Member.

4.1. Environmental Indicators and Paleo-Water Depth Estimates

Cluster analysis of the benthic foraminiferal assemblage indicates three biofacies groups in the Sunken Meadow Member (Figure 3a, Table 3). Group 1 (Table 3) contains the bottom nine samples and is characterized by Globocassidulina subglobosa, Epistominella danvillensis, and Rosalina bassleri as the predominate taxa (Table 3). Epistominella danvillensis is generally considered to be characteristic of water depths < 500 m [46]. Gibson [23] reported this species from the section at Lee Creek Mine, NC, and noted that abundances of >9% occur in areas of greater water depth. Globocassidulina subglobosa is a cosmopolitan infaunal species that has a wide geographic and bathymetric range, occurring from middle shelf to abyssal depths [47]. This species has been observed in water depths 80 to 140 m north of Cape Hatteras [48] and is generally considered to be an opportunistic species that feeds on phytodetritus [49]. Several studies have suggested that G. subglobosa flourishes in cooler, well-oxygenated bottom waters with high organic carbon flux [49,50,51]. Wilcoxon [52] found G. subglobosa to be diagnostic of modern outer shelf (61–183 m water depth) environments.

Group 2 contains two samples collected in a shell bed near the middle of the outcrop. This biofacies is characterized by two predominant taxa: Rosalina bassleri and Globocassidulina subglobosa (Table 2) and contains the highest abundance of Buliminella elegantissima and Bulimina elongata—generally considered to indicate lower oxygen environments [24,53]. Bulimina elongata is known to tolerate oxygen depletion and high organic influx [53]. Bulimina species occur in continental shelf environments with shallow oxygen minimum zones [53,54]. Buliminella elegantissima occurs in low abundance on modern continental shelves [48], but high numbers occur only in nutrient-rich waters or areas with high organic carbon content [24,53]. Katrosh and Snyder [53] posit that modern shelves with oxygen minima are associated with high organic production and upwelling.

Group 3 contains the four uppermost samples collected from the outcrop (Figure 3a) and is dominated by Rosalina bassleri, with Globocassidulina subglobosa and Epistominella danvillensis also occurring in relatively high abundance (Table 3). These samples contain the highest abundance of Hanzawaia concentrica, which is generally considered to be indicative of warmer, shallower waters [23,24,53]. Hanzawaia concentrica is a minor component of the total benthic assemblage at Spring Grove (Table 3), occurring in low abundance (<9%). This species generally occurs < 1% at depths of 220 m, but >5% at less than 100 m water depth in modern settings [48,53].

Miliolid taxa are absent in the Sunken Meadow Member at Spring Grove; these taxa are generally robust, thick-walled forms that are common in modern shallow shelf deposits [28,48,52]. The low abundance of other taxa generally considered to be diagnostic of shallow (<40 m) shelf environments (e.g., Elphidium excavatum, Cibicides lobatulus, Cibicides americanus) supports water depths greater than 40 m.

In modern environments, planktic foraminifera occur in low numbers in inner to middle shelf settings, rapidly increasing in abundance in outer shelf to slope environments [48,52,55,56]. Gibson [55] found that planktic foraminifera occur at very low values (0–1%) in shallow (30–40 m water depth) environments south of Cape Cod, MA, USA. Along the US Atlantic Coastal Plain, the percentage of planktic foraminifera generally reach values >20% in water depths greater than 30 m but less than 70 m [52,55,56]. Wilcoxon [52] found that planktic foraminifera occur in low density (<100 specimens per gram of sediment) at water depths generally less than 100 m along the southern part of the ACP, south of Cape Hatteras, NC. The low density and percent of planktic foraminifera indicate water depths between 30 and 100 m for the whole unit at Spring Grove.

The number of benthic species ranges from 22 to 31 throughout the unit (Table 2). This is consistent with water depths < 100 m for modern western Atlantic shelf settings [52]. The number of individual species typically decreases towards the shelf break [57]. Generally, Fisher’s alpha values greater than 5 (α > 5) indicate normal marine environments [57]. The benthic assemblage present at Spring Grove supports a middle to outer neritic, normal marine depositional environment.

At Spring Grove, the Sunken Meadow Member is composed of gravelly to slightly gravelly (shelly) fine quartz sands. The sands slightly coarsen upward, with a slight increase in the percentage of sand and a decrease in muds (Figure 2). The mean grain-size also supports an overall upwards coarsening (Figure 2). Bailey [44] suggested a maximum depositional water depth of 40 m for the Sunken Meadow Member in the Albemarle Embayment based on grain-size distribution and molluscan and foraminiferal fossil assemblages recorded by Gibson [23]. In modern settings, areas of gravelly sands are deposited on the middle (~100 m water depth) shelf south of Cape Hatteras, NC [52]. Much of the western Atlantic shelf in this area is covered by very fine to medium sands [52]. The grain-size data collected from Spring Grove, in combination with the benthic foraminiferal data, indicate water depths deeper than those proposed by Bailey [44].

The shift in species abundance from Biofacies 1 to Biofacies 2—a marked decrease in G. subglobosa and an increase in B. elongata and Bu. elegantissima—indicates a shift from well-oxygenated to oxygen-depleted bottom waters. Biofacies 3 marks a return or recovery to well-oxygenated bottom waters. The change in abundance of G. subglobosa coincides with a decrease in SSTs (Figure 4) and could indicate changes in surface water productivity. These shifts likely indicate localized, intermittent periods of upwelling in the Salisbury Embayment during the early Pliocene.

4.2. Bottom Water and Sea Surface Temperatures

Several studies have reconstructed bottom water temperatures (BWTs) during the early Pliocene. Ostracod-based paleotemperature reconstructions for the Sunken Meadow Member in southeastern Virginia indicate annual BWTs ranging from 12.5 °C (winter) to ~20 °C (summer) [42,45]. Cronin et al. [42] give BWTs for the Sunken Meadow Member of 10.1 °C (winter) and 12.1 °C (summer); modern BWTs at the same latitude (37° N) for winter and summer are ~5.5 °C and 14.5 °C, respectively [42]. Temperature estimates derived from molluscan-based stable isotopes indicate a summer BWT < 20 °C [58]. Snyder et al. [59] interpreted the benthic foraminiferal assemblage at Lee Creek Mine as representing a cooler temperate environment and proposed localized upwelling in the Albemarle Embayment. Modern seasonal BWTs on the shallow shelf (<100 m) around Cape Hatteras are ~15–25 °C (winter, [52]) and ~20–27 °C (summer, [11,52]).

Few studies have looked at sea surface temperatures (SST) for the Sunken Meadow Member. Dowsett and Wiggs [17] analyzed one sample that was dominated by the planktic species Globigerina bulloides. This species occurs in high numbers in water temperatures between 5 and 15 °C [17], suggesting cooler water deposition. The SST data presented here, derived from $U_{37}^{K^{\prime }}$, give higher values of ~26 °C to 28 °C (Figure 4). Modern SST values for Cape Charles, VA, are ~4 °C (winter) to ~25 °C (summer) [11]. Seasonality is not determined in this study, but if we assume the alkenone-derived SSTs for the Sunken Meadow Member are representative of summer temperatures, these data indicate temperatures ~2 °C higher than modern SSTs for southeast Virginia during the early Pliocene. This is consistent with previous studies that indicate the mid-latitude (34–37° N) western Atlantic SSTs were 2–3 °C higher than today [21].

5. Conclusions

The benthic foraminiferal, grain-size distribution, and alkenone sea surface temperature records suggest a middle to outer neritic depositional environment for the Sunken Meadow Member at Spring Grove, SE Virginia, similar to modern conditions found south of Cape Hatteras, NC [45,52]. Regional intermittent upwelling in the Salisbury Embayment during the early Pliocene is supported by a slight decrease in sea surface temperatures, combined with indications of both cool and warm water taxa. With increased global warmth, subtropical to warm temperate zonations are likely to expand poleward, with increased oceanic heat transfer from low to high latitudes. The small sample size for this study limits the depositional interpretation for the Sunken Meadow Member, though future work is planned for other exposures of this unit. The results of this study will be incorporated into a larger paleoecological analysis of the western Atlantic basin during the Pliocene.