Bioassessment of a Northwest Florida Estuary Using Benthic Macroinvertebrates

Abstract

Benthic invertebrate community composition was surveyed across the salinity gradient of the Pensacola Bay Estuary in Florida during summer 2016. Macrofauna densities ranged from 1000 to 9300 individuals m⁻², with highest densities occurring at the upper estuary and the lowest in the mid- and lower estuary. Taxonomic richness and Shannon diversity were lowest in the upper estuary and increased along the salinity gradient. Small-bodied, near-surface infaunal polychaete species (e.g., Mediomastus ambiseta and Paraprionospio alata) dominated the macrofaunal community in fine sediment areas. We calculated the Gulf of Mexico Benthic Index of Biological Integrity for each site and compared the index scores with those from Environmental Monitoring and Assessment Program - Estuaries, an earlier benthic assessment model. Condition evaluations by the different models did not match across all sites in this study; however, scores consistently indicated that most sites were at or near degraded levels, implying that Pensacola Bay represents a marginal habitat for a “healthy” benthic macrofauna community. This study provided new information about the benthic communities and sediments in the Pensacola Bay estuary.

INTRODUCTION

Benthic condition indices are tools used to assess aquatic habitat quality and aid in detection of degraded environments (Diaz et al. 2004). These indices use benthic invertebrate community attributes, including number of species, diversity, abundance, and relative abundance of major taxa, pollution tolerance, and trophic functional group. These attributes have been shown to covary with environmental stressors such as eutrophication, hypoxia, and sediment contamination (Engle and Summers 1999). Benthic condition indices combine these measures to allow complex environmental data to be meaningfully communicated to stakeholders, including the public. Originally developed for freshwater systems, the Benthic Index of Biological Integrity (B-IBI) provided a means of distinguishing among degraded and nondegraded conditions, accounting for variation in geographic region, salinity, and sediment type (Karr et al. 1986; Karr 1991). More recently, region-specific B-IBIs have been developed for estuaries and near-coastal waters of the United States (Weisberg et al. 1997; Engle and Summers 1999; Van Dolah et al. 1999; Paul et al. 2001; Llansó 2002). Regarding their utility, B-IBIs are used in more than 80% of United States state water quality management programs (Southerland and Stribling 1995; Norris and Hawkins 2000).

Specific to the northern Gulf of Mexico (GOM) estuarine and coastal waters, the GOM B-IBI (Tetra Tech 2011) was developed to improve upon the Environmental Monitoring and Assessment Program - Estuaries (EMAP-E) northern GOM benthic index (Engle et al. 1994; Engle and Summers 1999) by accounting for subregional differences in the expected range of ecological conditions (e.g., salinity) and, consequently, the macrobenthic community structure across the northern GOM. Using data from EPA’s National Coastal Assessment program (USEPA 2008, 2012), the GOM B-IBI was calibrated and validated using data from about 1300 randomized sites distributed across estuaries of Texas, Louisiana, Mississippi, Alabama, and Florida. As with other B-IBIs, the GOM B-IBI ranges from 0 (poor condition) to 100 (good condition). The threshold for poor condition was based on the 25th percentile of best available conditions in a particular latitude and salinity zone (Tetra Tech 2011).

Despite the considerable effort required to develop the GOM B-IBI, it has yet to be applied in any GOM estuary. Thus, 1 objective of this study was to apply the GOM B-IBI in the Pensacola Bay estuary using benthic macrofauna community data collected at sites spanning the salinity gradient, including several mesohaline sites that are known to be vulnerable to the development of seasonal bottom-water hypoxia. We measured sediment metal concentrations to evaluate the likelihood that metal toxicity was contributing to a degraded benthic condition. For comparison, we also calculated the EMAP-E index, which has been previously used to characterize benthic condition in the northern GOM (Engle and Summers 1999).

METHODS

Study area

The Pensacola Bay estuary (Figure 1) adjoins the northeastern GOM and is comprised of several subestuaries, including Escambia Bay, Blackwater Bay, East Bay, and Pensacola Bay proper. Pensacola Bay connects to the GOM through the narrow (800 m wide) pass at the western end of Santa Rosa Island (Hagy and Murrell 2007). This moderately sized (370 km2), shallow (mean depth: 3.4 m), microtidal (mean tidal range: 0.37 m) river-dominated ecosystem is situated within an 18 318 km2 watershed made up of pine forests, cropland, pastures, and urban landscapes (Schroeder and Wiseman 1999; Murrell et al. 2009). Pensacola Bay is strongly influenced by seasonal and episodic freshwater flows, which contribute to the system becoming strongly stratified and hypoxic during warmer summer months (Hagy and Murrell 2007). Pensacola Bay suffered from extensive industrial point-source pollution during the 1950s and 1960s, which has since been significantly reduced or eliminated.

Figure 1.

Location of sites sampled along the Escambia-Pensacola Bay estuarine axis during summer 2016. Inset denotes the location of the Pensacola Bay system within the Gulf of Mexico.

Sample collection and laboratory analysis

The sampling locations (Figure 1) were oriented along the estuarine axis to capture the salinity gradient and followed an established Pensacola Bay estuary survey design (Hagy et al. 2006; Hagy and Murrell 2007). Sediment samples were collected during June to July 2016 using a Wildco box-corer (surface area: 232 cm2). At each site, 3 grab samples were collected and sieved onboard using pumped surface-site water. The material on the 0.5 mm screen was collected and preserved in 5% buffered formalin-rose Bengal solution and transferred to 70% ethanol within 60 days.

At each site, near bottom (~0.2 m from the bottom) water temperature, salinity, and dissolved oxygen were measured using a Sea-Bird Model SBE 25 instrument. Additional sediment samples were collected to characterize bulk properties, including grain size, organic matter, calcium carbonate content, and trace metal concentrations. Grain size (percent sand, silt, clay) was measured using the sieve-pipette technique (Carver 1971; Folk 1980). Percent organic matter and calcium carbonate were measured as loss-on-ignition at 550 °C and 1000 °C, respectively (Percival and Lindsay 1997; Heiri et al. 2001). Trace metals included Ag, As, Be, Cd, Cr, Cu, Pb, Ni, Sb, Se, Tl, and Zn, and were analyzed using inductively coupled plasma-mass spectrometry (USEPA 2014a). Of these 12 metals, 8 have been assigned Effects-Range-Low (ER-L) and Effects-Range-Median (ER-M) threshold values (Long et al. 1995).

Macrobenthic analysis and index calculations

For macrofauna community analysis, the macrofauna samples collected from each site were pooled to increase area sampled and reduce spatial variability within site. Total abundance and species composition were determined from pooled samples. Macrofauna were identified to the lowest practical taxonomic level, usually genus or species. Both sorting effectiveness (90% complete) and identification accuracy (10% sample re-identification, 90% similarity) were measured to assess the quality assurance. Species nomenclature was standardized against the World Register of Marine Species database (WoRMS Editorial Board 2019).

Each taxon encountered was classified by ecological functional role and trophic group as determined from published descriptions, inferences from phylogenetic and taxonomic relationships, or by best professional judgment (Tenore et al. 2006; Tetra Tech 2011). Relative tolerance to stress, including organic enrichment and hypoxia stress, was estimated for more than 1500 benthic macroinvertebrate taxa from estuarine and near-coastal waters of the GOM using results from exposure studies, interspecies extrapolated models, and best professional judgment determinations (Tetra Tech 2011). Each taxon was assigned a stressor tolerance category value ranging from 1 (sensitive) to 5 (tolerant).

Community statistics (e.g., taxa richness, Pielou’s evenness, Shannon diversity) were calculated using PRIMER-E v. 7 (Clarke and Gorley 2015). Taxa richness and stressor tolerance category ratings were used to calculate Beck’s biotic index scores for each site (Beck 1955). The Beck’s biotic index, a weighted measure of taxa richness for taxa sensitive to stress (i.e., taxa assigned to stressor tolerance categories 1 or 2), was calculated by counting taxa assigned to stressor tolerance category 1 twice and those in category 2 once (e.g., Beck’s biotic index = 2 [n Category 1] + [n Category 2], where n = number of taxa).

Following GOM B-IBI guidelines, the sites were classified as low salinity (LS) (0.5–18) or high salinity (HS) (18–40) based on near-bottom water salinity measured during sampling. While the individual component metric calculations differed slightly among LS and HS index classes, all were based on various taxonomic traits and ecological function measures (Table 1). Component metrics for each index class were calculated and averaged to yield the composite index values, which were then compared with class-specific index thresholds (e.g., LS: 64.7; HS-SF: 58.4; HS-xF: 57.1) to evaluate site condition.

Table 1.

Gulf of Mexico Benthic Index of Biological Integrity component metrics and scoring formulas for low salinity, high-salinity south Florida, and high-salinity non-Florida classes

Index class-metric	Component metric	Scoring formula
LS-1	Percent individuals as Bivalvia	100 × (96.0 – metric value)/96.0
LS-2	Percent individuals as Spionidae	100 × (58.9 – metric value)/58.9
LS-3	Percent individuals as predators	100 × metric value/76.0
LS-4	Percent individuals as tolerant	100 × (58.8 – metric value)/58.8
LS-5	Beck’s biotic index	100 × metric value/8.5
HS-SF-1	Number of taxa as Bivalvia	100 × metric value/6.1
HS-SF-2	Percent individuals as Polychaeta	100 × (78.1 – metric value)/75.3
HS-SF-3	Percent individuals as interface feeders	100 × (metric value – 17.2)/60.5
HS-SF-4	Percent individuals as tolerant	100 × (32.8 – metric value)/31.8
HS-SF-5	Beck’s biotic index	100 × (metric value – 1)/11.0
HS-xF-1	Percent individuals as Spionidae	100 × (70.3 – metric value)/70.3
HS-xF-2	Percent taxa as Polychaeta	100 × (90.0 – metric value)/65.0
HS-xF-3	Percent individuals as intolerant	100 × metric value/48.4
HS-xF-4	Percent individuals as tolerant	100 × (84.5 – metric value)/84.5
HS-xF-5	Percent taxa as intolerant	100 × metric value/33.3

LS = low salinity; HS-SF = high-salinity south Florida; HS-xF = high-salinity non-Florida.

Table contains information reported by Tetra Tech (2011).

Two HS index classes were established in the GOM B-IBI: 1 for Florida south of latitude 28.0°N (HS-SF), and the other encompassing the other Gulf states, excluding Florida (HS-xF). Although the Florida panhandle region, which includes Pensacola Bay, was found to have characteristics distinct from the other regions, insufficient data were available to generate an applicable panhandle region-specific index. Both HS index classes were applied to the data and are reported here.

EMAP-E benthic index scores were calculated for each site as described in Engle and Summers (1999) and reported within the context of a Gulf-wide reference dataset (from Florida Bay, FL to Laguna Madre, TX) collected from 2000 to 2006 (USEPA 2008, 2012) and spanned tidal freshwater to marine habitats. Index scores were normalized to a range of 0 to 100 to facilitate comparison with the GOM B-IBI scores and were calculated using the following formula: = (1.5710 × proportion of expected Shannon diversity) + (−1.0335 × mean abundance of tubificid oligochaetes) + (−0.5607 × percent capitellid polychaetes) + (−0.4470 × percent bivalves) + (0.5023 × percent amphipods).

The proportion of expected diversity was calculated by dividing the observed diversity value by an expected value based on the reference dataset, and was included to account for the effects of salinity on benthic community structure. Scores below 30 were deemed degraded and those above 50 were considered in good condition.

Sediment contamination was evaluated using National Coastal Condition Report (NCCR) Sediment Quality Index methodology (Long et al. 1995; USEPA 1999, 2001, 2004, 2008, 2012), which considers metals that are known to adversely affect aquatic organisms. Thresholds known as the ER-L and ER-M were derived from laboratory toxicological studies and represent the concentrations expected to cause toxic effects approximately 10% and 50% of the time, respectively. Sites were deemed to have “good” sediment quality if no metal exceeded the ER-M or if fewer than 5 metals exceeded the ER-L (USEPA 2012). This sediment chemistry information was compared with the GOM B-IBI scores to identify any associations between degraded benthic communities and sediment contaminants.

RESULTS

Sediment conditions

Results from granulometry analysis showed that sediments were composed primarily of muds and clays, with sand making up 10% or more of the sediment only in the most upper and lower Bay sites and the midBay shoal–ward transect stations (Table 2). Sediments were organic rich (10%–15%), peaking at 15% in the lower Bay (e.g., P08). Of the 12 trace metals measured in sediments, 8 have been assigned ER-L and ER-M threshold values. No metal concentrations measured exceeded the ER-M thresholds; however, some As (8.2 ppm) and Ni (20.9 ppm) concentrations exceeded ER-L thresholds (Table 3). Since fewer than 5 metals exceeded ER-L thresholds, sediment quality across all of the sites was considered “good” in terms of metals toxicity using the USEPA NCCR Sediment Quality Index evaluation methodology (USEPA 2014b).

Table 2.

Bottom water and sediment characteristics measured

Site	Depth (m)	Salinity	DO (mg/L)	% Sand	% Silt	% Clay	% Organic matter	% Calcium carbonate	Folk (1980) classification
P02	1.9	10.2	5.4	10.5	39.2	50.3	9.3	4.7	sM
P03	3.7	7.4	5.5	9.9	32.4	57.7	12.3	6.4	M
P04	2.7	15.3	5.0	4.3	28.1	67.6	11.5	6.3	C
P05	4.0	26.2	3.3	0.8	26.4	72.8	14.2	8.4	C
P06	5.5	22.3	3.1	0.3	13.0	86.7	14.4	9.9	C
P07	10.1	33.4	5.4	2.3	21.4	76.3	14.7	15.5	C
P08	10.9	34.3	5.3	37.8	17.8	44.4	15.3	18.1	sC
P05-mid	3.4	30.6	6.1	29.8	0.9	69.3	8.7	6.1	sC
P05-east	3.0	25.2	4.7	76.0	0	24.0	13.1	13.2	cS

DO = dissolved oxygen; M = mud; C = clay; sM = sandy mud; sC = sandy clay; cS = clayey sand.

Table 3.

Trace metal concentrations (ppm) measured

	Site								Thresholds
Metal	P02	P03	P04	P05	P06	P07	P08	Mean	ER-L	ER-M	Historic mean/maximum
Arsenic	32	28	36	30	26	20	20	27	8.2	70	na
Cadmium	0.97	0.72	0.82	0.87	1.0	1.1	0.89	0.91	1.2	9.6	<1/nd
Chromium	74	57	79	77	76	54	44	66	81	370	39.7/110
Copper	22	15	18	19	19	17	14	18	34	270	8.7/43
Lead	32	26	36	37	38	26	17	30	46.7	218	18.5/43
Nickel	22	16	23	22	22	17	14	19	20.9	51.6	8.8/19
Silver	0.12	0.076	0.082	0.11	0.14	0.10	0.085	0.10	1.0	3.7	na
Zinc	130	100	130	120	120	79	54	105	150	410	43.2/98
Antimony	0.58	0.43	0.49	0.52	0.62	0.66	0.53	0.55	na	na	na
Beryllium	2.5	2.0	2.7	2.4	2.1	1.4	1.1	2.0	na	na	na
Selenium	1.5	0.94	1.3	1.3	0.89	0.66	0.96	1.1	na	na	na
Thallium	0.44	0.32	0.37	0.39	0.47	0.50	0.40	0.41	na	na	na

For comparison, the Effects-Range-Low (ER-L) and Effects-Range-Median (ER-M) threshold values (Long et al. 1995) and historic mean and maximum concentrations from Olinger et al. (1975) are included. Values exceeding the ER-L threshold are shown in bold italics.

na = not available; nd = not detected.

Macrobenthic community composition

A total of 53 unique macrofauna taxa was identified in this study. Annelids accounted for 27 species and contributed 87% of total macrobenthos abundance. Nine mollusc species were identified, contributing 4% of total abundance. Eight crustacean taxa accounted for 2% of total abundance. The remaining 9 taxa included nemerteans, sipunculids, phoronids, and echinoderms, and they made up 7% of total macrobenthos abundance. Taxa richness was highest at the lower Bay sites (P07 and P08) (Table 4). Macrofauna abundance ranged from roughly 1000 to 9300 individuals m−2, with highest abundance at LS sites in the upper Bay and lowest abundance at mid- and lower Bay sites (Table 4). Shannon diversity values ranged from 0.7 to 2.6, with the lowest values in the upper Bay and increasing toward Pensacola Pass (Table 4).

Table 4.

Benthic macrofauna community parameters calculated from each site sampled

Site	Macrofauna abundance (m−2)	Taxa richness	Shannon diversity (H’)	Pielou’s evenness (J’)	Beck’s biotic index	% Abundance of stress-sensitive taxa	% Abundance of top 2 dominant taxa	Two most numerically dominant taxa (abundance m−2)
P02	3558	11	0.748	0.312	2	9.6	91.9	Polychaeta, Capitellidae: Mediomastus ambiseta (2941), Spionidae: Streblospio benedicti (330)
P03	9268	17	0.724	0.256	4	11.6	93.5	Polychaeta, Capitellidae: Mediomastus ambiseta (7704), Spionidae: Streblospio benedicti (961)
P04	2425	16	1.445	0.522	2	17.2	78.1	Polychaeta, Capitellidae: Mediomastus ambiseta (1550), Spionidae: Streblospio benedicti (344)
P05	1636	16	2.195	0.792	6	23.7	43.9	Polychaeta, Capitellidae: Mediomastus ambiseta (430), Cossuridae: Cossura sp. (287)
P06	1090	13	2.146	0.837	3	14.1	47.4	Polychaeta, Spionidae: Paraprionospio alata (301), Cossuridae: Cossura sp. (215)
P07	1277	19	2.571	0.873	4	11.2	33.7	Polychaeta, Spionidae: Paraprionospio alata (273), Capitellidae: Mediomastus ambiseta (158)
P08	2066	27	2.625	0.796	4	33.3	41.0	Polychaeta, Spionidae: Paraprionospio alata (445); Oligochaeta (402)
P05-mid	2583	19	2.112	0.717	4	10.0	52.8	Phoronida: Phoronis architecta (875); Polychaeta, Capitellidae: Mediomastus ambiseta (488)
P05-east	3429	17	1.336	0.471	4	11.7	76.2	Polychaeta, Capitellidae: Mediomastus ambiseta (2410), Chaetopteridae: Spiochaetopterus costarum (201)

Stress-sensitive taxa included those classified in stressor tolerance categories 1 or 2, based on the values assigned by Tetra Tech (2011).

Two taxa (the gastropod Acteocina canaliculata and a Nuculanid bivalve), classified as being the most sensitive to stress (stressor tolerance category 1) were identified during this study, and only the midBay sites (P05, P06) had at least 1 of these species present. There were 9 taxa collected during this study that were classified in stressor tolerance category 2 (polychaetes Phyllodoce sp., Sigambra sp., Spiochaetopterus costarum, Sthenelais boa, and Streblospio benedicti; bivalves Corbula sp., Macoma sp., cumacean Oxyurostylis smithi; and isopod Edotia triloba), and all study sites had at least 1 of these taxa present, with some of these species being the second-most numerically dominant taxa at upper and midBay sites (e.g., P02, P03, P04, and P05-east (Tables 4 and 5). Stress-sensitive taxa made up approximately 10% to 33% of the total abundance of macrofaunal organisms across the sites (Table 4), with the greatest proportion being at P08.

Table 5.

Abundance (individuals m−2) of the 5 dominant taxa recovered from each site

Taxa	Tolerance category	Functional group	P02	P03	P04	P05	P06	P07	P08	P05-mid	P05-east
Acteocina canaliculata	1	pred									72
Aricidea sp.	3	infc						129
Boonea bisuturalis	3	pred			43
Carinoma sp.	3	pred	57	43
Carinomella lactea	3	pred		115	57		57				115
Cossura sp.	4	infc			57	287	215	86		416
Glycinde multidens	3	pred	29	57	43						172
Hermundura sp.	3	pred	29
Leitoscoloplos robustus	3	subsrf	29	115	72
Leucon sp.	unk	infc							115
Lineidae	3	pred							72
Mediomastus ambiseta	3	subsrf	2941	7704	1549	430	158	158		488	2410
Mulinia lateralis	3	infc	43				100	86		115
Oligochaeta	3	subsrf							402
Paraonidae	4	infc						115
Paraprionospio alata	2.5	infc			43	258	301	273	445	129	72
Phoronis architecta	3	infc								875
Podarkeopsis levifuscina	4	pred		43
Sagitta sp.	unk	unas					57		273
Sigambra tentaculata	2	pred		57		100	57		115
Spiochaetopterus costarum	2	infc			57						201
Streblospio benedicti	2	infc	330	961	344	201
Tellinidae	3	infc	57		43				72		72

Tolerance category is a relative tolerance to stress for a given taxon and is based on values assigned by Tetra Tech (2011), with 1 representing the most sensitive and 5 the most tolerant to stress.

Functional group represents the ecological functional role and trophic group as determined from published descriptions, inferences from phylogenetic/taxonomic relationships, or by best professional judgment (Tenore et al. 2006; Tetra Tech 2011).

unk = unknown (no value assigned); infc = interface and water column feeder; pred = predator; subsrf = subsurface feeder; unas = unassigned.

Index scores

Benthic condition evaluations estimated by the different benthic index models did not agree across all sites sampled in this study; however, the condition of at least half of the sites was estimated to be at or near degraded levels across all indices (Figure 2). The GOM B-IBI index classified 66% of the LS sites and, depending on which index class was applied, 66% to 100% (HS-xF and HS-SF, respectively) of the HS sites as degraded because the composite B-IBI score did not exceed the 25th percentile threshold value for scores obtained from sites considered to be in good condition. The EMAP-E benthic index calculations characterized 44% of the sites in good condition, 44% in degraded condition, and 11% (1 site, P04) in marginal condition (Figure 2).

Figure 2.

Benthic index scores for the Pensacola Bay study sites (bars) and estimated degradation thresholds (horizontal lines). Included are scores calculated using (A) GOM B-IBI HS-SF index, (B) GOM B-IBI LS index (hatched bars and dotted threshold) for low-salinity sites and GOM B-IBI HS-xF (solid bars) for high-salinity sites, and (C) EMAP-E benthic index, including thresholds for good, marginal, or degraded condition ratings (horizontal lines). The shaded boxes on the right are included as visual aids indicating threshold designations. GOM = Gulf of Mexico; B-IBI = Benthic Index of Biological Integrity; HS-SF = high-salinity-south Florida; LS = low salinity; HS-xF = high-salinity-non-Florida; EMAP-E = Environmental Monitoring and Assessment Program - Estuaries.

The 2 LS sites with GOM B-IBI scores below the good condition threshold level for the LS index class (P02 and P04) were considered biologically degraded (Figure 2 and Table 6). These sites both had lower Beck’s biotic index scores compared with P03, indicating that the poorer scoring sites had a lower proportion of stress-tolerant taxa present. The EMAP-E scores for all 3 of the LS sites fell below the good threshold and 2 (P02 and P03) fell below the marginal condition threshold and were considered to be in degraded condition (Figure 2).

Table 6.

Gulf of Mexico Benthic Index of Biological Integrity component metric scores (calculated using formulas in Table 1) and the combined index score for each study site using (A) the low salinity, (B) high-salinity south Florida, and (C) high-salinity non-Florida index classes

A
	GOM B-IBI component metrics					Index Score
Site	LS-1	LS-2	LS-3	LS-4	LS-5
P02	98.7	84.3	5.3	100.0	23.5	62.4
P03	99.7	82.1	5.5	99.2	47.1	66.7
P04	98.2	72.9	12.5	93.0	23.5	60.0

	B
Site	HS-SF-1	HS-SF-2	HS-SF-3	HS-SF-4	HS-SF-5	Index score
P05	2.0	90.4	46.5	21.1	6.0	32.7
P06	2.0	75.0	61.8	22.4	3.0	32.3
P07	1.0	80.9	67.4	23.6	4.0	31.1
P08	4.0	42.4	45.1	6.3	4.0	54.0
P05-mid	1.0	53.9	65.6	17.8	4.0	40.6
P05-east	2.0	90.8	15.5	2.1	7.0	36.2

	C
Site	HS-xF-1	HS-xF-2	HS-xF-3	HS-xF-4	HS-xF-5	Index score
P05	28.1	56.3	21.9	21.1	25.0	61.5
P06	27.6	61.5	6.6	22.4	15.4	47.6
P07	24.7	68.4	9.0	23.6	21.1	50.4
P08	24.3	40.7	11.1	6.3	14.8	60.3
P05-mid	7.8	52.6	6.1	17.8	15.8	57.1
P05-east	3.8	58.8	11.3	2.1	35.3	73.9

The LS metrics were applied to sites with bottom water salinities between 0.5 and 18, and HS metrics were applied to sites with bottom water salinities of 18 to 40. GOM B-IBI scores in bold italics are considered degraded (i.e., falling below the 25th percentile GOM B-IBI threshold value for each index class; e.g., LS: 64.7; HS-SF: 58.4; HS-xF: 57.1).

GOM = Gulf of Mexico; B-IBI = benthic index of biological integrity; LS = low salinity; HS-SF = high-salinity south Florida; HS-xF = high-salinity non-Florida.

The GOM B-IBI HS-SF composite scores for the 6 HS sites in this study (P05, P05-mid, P-05 east, P06, P07, and P08) ranged from 31.1 to 54.0 of a possible100, and all scores fell below the HS-SF good-condition threshold value (58.4) (Figure 2 and Table 6). Station P08 had the lowest proportion of organisms that were polychaetes and the greatest number of bivalve taxa compared with the other HS sites, and yielded the highest HS-SF composite index score of the 6 HS sites in this study. Stations P05, P06, and P07 scored on the lower end of the observed HS-SF composite index score range. The benthic community from these stations were numerically dominated by polychaetes (>75% of all individuals collected) and >20% of all of the individuals collected from each of these stations were species considered to be stress tolerant, that is, assigned a stressor tolerance category value of 4 or 5 by Tetra Tech (2011).

The GOM B-IBI HS-xF composite scores were higher than the corresponding HS-SF model composite index score for each site (Figure 2). Only sites P06 and P07 fell below the HS-xF good condition threshold (57.1). Polychaete species were the dominant taxa, comprising >60% of all the taxa recovered from these stations, and >20% of all individual macrofaunal organisms were categorized as being stress-tolerant species (Tetra Tech 2011).

The EMAP-E benthic index scores, scaled from 0 to 100, ranged from 4 to 64 across the study area, with most midBay stations scoring in the good condition range and stations P02, P03, P08, and P05-east yielding the lowest scores observed (Figure 2). The low scores that were observed resulted from the presence of tubificid oligochaetes, which were only recovered in the samples collected from station P08, or the high proportional abundance (>70%) of Capitellid polychaetes at stations P02, P03, and P05-east. Amphipods, one of the benthic groups included in the EMAP-E index calculations and indicative of good habitat condition, were not present at any of the stations.

DISCUSSION

Benthic indices have been widely used to characterize benthic community health and, by extension, the environmental quality of coastal and estuarine systems (Weisberg et al. 1997; Engle and Summers 1999; Van Dolah et al. 1999; Paul et al. 2001; Llansó 2002; Diaz et al. 2004). Theses indices typically evaluate condition through measurements of various benthic macroinvertebrate community parameters and compare those with reference values representative of pristine conditions (Llansó et al. 2003). It can be difficult to attribute changes observed in biological indicators to a natural variability in physical condition, such as changes in salinity, or to anthropogenic effects, such as sediment contamination or eutrophication. The GOM B-IBI was developed to include a site classification system to help partition natural variability (e.g., salinity and latitude) and, thus, increase sensitivity for detecting changes in ecological condition due to anthropogenic stress. Despite the considerable effort invested to develop this regionally specific benthic condition index, this study was the first application of the GOM B-IBI to this region of the GOM coastline. We applied 2 indices, the GOM B-IBI and EMAP-E benthic index, to our data with the goal of evaluating the status of the benthic communities in Pensacola Bay. However, we found that our ability to draw conclusions was hindered by disagreement in ecological condition status across the study area using the different index approaches. This pointed to a need for further development and testing of a benthic index for the Florida panhandle.

Two HS index classes were established in the GOM B-IBI: one for Florida south of latitude 28.0°N (HS-SF), and the other encompassing the other Gulf states, excluding Florida (HS-xF). The Florida panhandle region, which includes Pensacola Bay, was found to have characteristics distinct from the other regions, and sites from the Florida panhandle were excluded during the development of the B-IBI due to insufficient data. Until data were available to generate a Florida panhandle region-specific index class, Tetra Tech (2011) recommended applying the HS-SF index to the panhandle as an interim solution. This was because the number of taxa identified from available reference sites in northwest Florida (10 taxa) was closer to that for south Florida (22 taxa) than to other Gulf states (73 taxa). Taxa richness measured across the HS sites in Pensacola Bay during this study was 48, approximately a median value between the taxa richness values measured from south Florida and the other Gulf states during index development, which may have contributed to the inconsistency in final scores when applying the HS-xF and HS-SF models to the HS sites. This inconsistency is problematic when trying to infer a benthic habitat condition using this index; therefore, conclusions as to benthic conditions using the GOM B-IBI, especially for the HS sites, should be viewed with caution. Until this can be resolved, the greatest utility of GOM B-IBI index may be in its application to evaluate trends by examining how index scores at a given location may change with time, which was beyond the scope of this current study, but worthy of future evaluation.

The GOM B-IBI scores did not coincide meaningfully with EMAP-E scores (Figure 2). The GOM B-IBI index suggested near degraded levels across much of the estuary, implying that environmental conditions in the system are not conducive to supporting a richer, more diverse macrofauna community. In contrast, the EMAP-E index graded much of the middle and lower estuary as being in marginal or good condition. Stratification-driven seasonal hypoxia and HS variability are well documented in the middle reaches of the Pensacola Bay estuary (Olinger et al. 1975; Hagy and Murrell 2007; Caffrey and Murrell 2016), but they are not common at the lower Bay sites (P07, P08). Lower GOM B-IBI scores resulted where hypoxia is commonly observed (Figures 2 and 3), yet the EMAP-E scores graded these sites as being in marginal or good condition. Furthermore, the EMAP-E index classified P08 and P05-east as degraded and P07 as good condition.

Figure 3.

Cross-sectional view of the Escambia-Pensacola Bay transect showing the probability of hypoxia (DO < 2 mg L−1) occurring in bottom waters during warm months (May–October). Probabilities were based on water column DO concentration profiles from multiple years (n = 64 dates) from 2002 to 2016. Location of sampling sites are represented by black circles situated at the sediment–water interface. The shading indicates the probability of hypoxia being observed (see scale bar); the dotted line shows the 0.1 probability contour. DO = dissolved oxygen.

Patterns in the GOM B-IBI scores were inconsistent with patterns observed in benthic macrofauna community parameters, including species composition, richness, or evenness. For example, regardless of index class applied (e.g., LS+HS-SF or LS+HS-xF), as the GOM B-IBI index score increased, diversity and evenness decreased, while dominance increased (Table 4). In contrast, increased EMAP-E index scores coincided with increased diversity and evenness and decreased dominance. This was true even with the P08 outlier included. P08 had high numbers of oligochaetes and spionid polychaetes, perhaps indicating a disturbance or stress at this site impacting the benthos. This site had a low probability of seasonal hypoxic conditions occuring there (Figure 3) (Olinger et al. 1975; Hagy and Murrell 2007; Caffrey and Murrell 2016), so some other disturbance may have contributed to the low to marginal index scores and benthic community conditions observed there.

The area around P08 has been repeatedly dredged, with as much as 3 m of sediment removed (USDON 1986), to maintain the turning basin for Naval vessels at the United States Naval Air Station-Pensacola. Although the most recently known large-scale dredging activities at P08 occurred nearly 3 decades prior to this study, physical disturbance and repeated resuspension of sediments in the turning basin from ongoing shipping activities may have ongoing effects on benthic communities (Newell et al. 1998; Schaffner et al. 2001; Wilber et al. 2005; Hinchey et al. 2006). While estuarine muds and sands are known to generally be rapidly colonized after initial dredging activities, repeated and frequent disturbances may restrict the community composition to being dominated by more opportunistic species rather than longer-lived and slow-growing equilibrium species (Newell et al. 1998). The macrofaunal community at P08 was dominated by species considered to be opportunistic and early colonizers following a disturbance, and included oligochaetes, a taxon that was only observed at this site. While metal concentrations observed were generally below levels considered detrimental for benthic fauna across all of the sites, the lowest concentrations were observed at P08. This result was somewhat surprising given the site’s proximity to the industrialized waterfront areas of downtown Pensacola and Naval Air Station-Pensacola, where the use of antifouling paints on ships and infrastructure is common. Furthermore, sediment contamination from metals was expected because this site was also located near the mouth of Bayou Chico, a heavily industrialized urban embayment generally considered to be the most polluted of the 3 urban bayous in the Pensacola area, with known high accumulated levels of sediment contaminants, including trace metals, polycyclic aromatic hydrocarbons (PAHs), pentachlorophenol, dioxins/furans, and polychlorinated biphenyls (PCBs) (Mohrherr et al. 2006). If we had measured PAHs or other organic contaminants instead of just metals in this study, perhaps a signal would have been evident.

It is unlikely that sediment metal toxicity is an important factor structuring benthic communities in Pensacola Bay. Of the 12 trace metals examined, only 2 exceeded published ER-L threshold values. The concentration of arsenic exceeded the ER-L at all sites sampled, and 4 sites had Ni concentrations exceeding the ER-L values (Table 2). Of the 4 sites with high Ni concentrations (P02, P04, P05, and P06), 3 had consistently low GOM B-IBI scores (Table 6). This may indicate that metals may have been among the stressors influencing the benthic community composition, however not likely toxic and exclusively driving the results observed. Although not measured during this study, PCBs are known to have persisted in the surface sediments of Escambia Bay for at least 50 y (Mohrherr et al. 2012) and may contribute to the patterns in benthic community distribution observed. Surface sediments from regions in upper Escambia Bay, in the vicinity of our stations P02 and P03, had PCB concentrations exceeding the Florida Department of Environmental Protection’s threshold effects level of 21.6 mg/kg (Mohrherr et al. 2012).

Macrofauna abundance extremes (i.e., high abundance or low abundance) may suggest a stressed community (Pearson and Rosenberg 1978). An increase in abundance can be expected at polluted sites with organic enrichment, whereas a decrease in abundance can be expected at sites with high pollution stress. In this study, macrofauna densities were greatest at the upper Bay sites (P02 and P03) and lowest at sites P06 and P07 (Table 4), with highest proportional abundance of stress-sensitive taxa, including the polychaetes Paraprionospio alba, Streblospio benedicti, and Sigambra tentaculata, occurring at sites P05 and P08 (Tables 4 and 5). Small-bodied infaunal polychaetes were among the most abundant taxa across the estuary, with Paraprionospio alata and Cossura sp. most predominant at the mid- and lower Bay sites and Mediomastus ambiseta and Streblospio benedecti dominating the upper Bay sites (Tables 4 and 5). These species are all subsurface or interface and water-column feeders (Tenore et al. 2006; Tetra Tech 2011) and are opportunistic; thus, they are often early colonizers following a disturbance (Pearson and Rosenberg 1978). Since a community that is rich with many types of organisms with even species abundance distributions is considered healthier and more ecologically stable compared with one with few types of organisms, and the taxonomic richness and evenness observed at the upper Bay sites was nearly half of what was observed for the lower Bay sites (Table 3), the lower Bay sites may be considered more healthy, stable environments. Except the low-scoring site P08, this pattern was reflected in the EMAP-E index scores, with the greatest index values measured at the mid- and lower Bay sites.

This study provides an uncommon and valuable update to the literature describing benthic communities in Escambia and Pensacola Bay, spanning a gradient of both salinity and average exposure to seasonal hypoxia. The application of established benthic indices using study data was somewhat inconclusive. The disagreement in results between the different indices prevented an unequivocal conclusion supporting the application of one index approach versus the other in Pensacola Bay. These results suggest that the new GOM B-IBI index may not be suitable for assessing the Pensacola Bay system. Until the GOM B-IBI can be refined to include panhandle-specific metrics, the EMAP-E index may be the most appropriate for assessing the region.

CONCLUSIONS

The results of this study suggest that more systematic testing and reference site assessments are needed to make the GOM B-IBI more useful for environmental management applications, particularly for HS regions of northern Florida estuaries with muddy sediments. Providing the additional information needed to more appropriately apply the GOM B-IBI to northwest Florida would result in valuable improvements to benthic indices applicable to this region. Because benthic community analyses are often expensive, the GOM B-IBI is already useful because it may be applied to a small number of sites within a system, if the appropriate benthic community data are available. Improving its applicability to northwest Florida estuaries would provide improved and more cost-effective guidance to inform environmental decision making within the broader geographic region.