Acute Impacts of Hurricane Ian on Benthic Habitats, Water Quality, and Microbial Community Composition on the Southwest Florida Shelf

Abstract

Tropical cyclones can severely disturb shallow, continental shelf ecosystems, affecting habitat structure, diversity, and ecosystem services. This study examines the impacts of Hurricane Ian on the Southwest Florida Shelf by assessing water quality, substrate type, and epibenthic and microbial community characteristics at eight sites (3 to 20 m in depth) before and after Ian’s passage in 2022. Hurricane Ian drastically changed substrate type and biotic cover, scouring away epibenthos and/or burying hard substrates in mud and sand, especially at mid depth (10 m) sites (92–98% loss). Following Hurricane Ian, the greatest losses were observed in fleshy macroalgae (58%), calcareous green algae (100%), seagrass (100%), sessile invertebrates (77%), and stony coral communities (71%), while soft coral (17%) and sponge communities (45%) were more resistant. After Ian, turbidity, chromophoric dissolved organic matter, and dissolved inorganic nitrogen and phosphorus increased at most sites, while total nitrogen, total phosphorus, and silica decreased. Microbial communities changed significantly post Ian, with estuary-associated taxa expanding further offshore. The results show that the shelf ecosystem is highly susceptible to disturbances from waves, deposition and erosion, and water quality changes caused by mixing and coastal discharge. More routine monitoring of this environment is necessary to understand the long-term patterns of these disturbances, their interactions, and how they influence the resilience and recovery processes of shelf ecosystems.

1. Introduction

Continental shelves are among the most biologically productive, diverse, and economically important ecosystems in the world, supporting most of the world’s marine biodiversity and fisheries harvests despite comprising less than 10% of total ocean area [1]. Because of their shallow nature and proximity to populated coastlines, these ecosystems are subject to an array of natural and anthropogenic disturbances. Among the most influential natural disturbances to continental shelves are major tropical cyclones, which alter the shelf communities via scouring, sediment transport, and changes in water quality [2,3,4,5]. Sediment movement during major cyclones can reveal new areas of hard bottom for benthic organisms to settle and can simultaneously bury other habitats with up to 1 m of unconsolidated sediments [6,7,8]. Physical (e.g., mixing, cooling), chemical (e.g., nutrient and organic carbon addition, dissolved organic matter remineralization), and biological (e.g., increased primary production) effects of tropical cyclones on shallow continental shelves can alter both the water column and benthic community composition [9,10,11,12,13]. However, the severity of these effects depends on several factors such as the cyclone’s intensity, trajectory and location, composition of the impacted marine community, types and amounts of allochthonous materials added, and resilience and response of the affected marine communities [14,15,16,17].

The pre- and post-typhoon impacts of Typhoon Maria on the pelagic ecosystem in the southern East China Sea were investigated and it was reported that the storm-induced physical disturbances, such as water column mixing, enhanced nutrient availability, which in turn led to higher primary production and increased chlorophyll-a concentrations in the shallow water column [13]. That study also reported that increased community respiration following the storm contributed to elevated pCO₂, highlighting the complex interactions between tropical cyclones and marine ecosystems and emphasizing the need for further research to fully understand the ecological consequences of such disturbances [13]. Similarly, Hurricane Harvey in the Gulf of Mexico altered phytoplankton productivity, thereby affecting the entire marine ecosystem [9]. While the effects of tropical cyclones on the water column can last for months [9,18], the physical damage to benthic habitats can have long-lasting impacts.

It has been reported that the ecological succession within the shelf environments disturbed by tropical cyclones can take up to 10 years to return communities to their pre-disturbance states [8]. Spatially heterogeneous and intermediate-intensity disturbances can prevent competitive exclusion and support biodiversity [19,20,21,22], whereas severe, large-scale disturbances may lead to long-term loss of habitat, biodiversity, and ecosystem functioning, especially in areas impacted by multiple anthropogenic stressors [23]. In shelf communities, where resilience has been compromised by anthropogenic stresses, storm-disturbed benthic habitats may never recover, and instead shift to alternate states [24]. Furthermore, anthropogenic climate change is predicted to increase the intensity of tropical cyclones in many regions of the world [25,26], which is likely to lead to harsher disturbances to vulnerable environments such as the Southwest Florida Shelf (SWFS).

Cyclone-induced benthic damage can alter marine community structure, productivity, and diversity, leading to cascading effects on water column communities and higher trophic levels [27]. While recent studies have highlighted the effects of tropical cyclones on surface and upper water column communities [11,12,13], the effects of cyclones on benthic habitats can be more severe and are less understood. Therefore, it is vital to explore how major cyclones affect benthic habitats in shallow continental margins like the SWFS. This study addresses a key gap in understanding how tropical cyclones influence benthic habitats and community structure, ultimately affecting food web interactions and fisheries production in continental shelves.

The SWFS encompasses a vast expanse of the Gulf of Mexico (GOM), yet it has only recently begun to be mapped and studied [28,29]. The synergy of freshwater discharge (e.g., Caloosahatchee River, the western Everglades, and submarine springs), shallow depths, and long residence times make the SWFS particularly susceptible to the “estuarization of the shelf” [30], i.e., decreased salinity and optical water quality, increased turbidity and sediment deposition, harmful algal blooms (HABs), hypoxia, and anoxia [31,32,33]. Tropical cyclone disturbances can contribute to HABs by introducing massive nutrient pulses to the shelf, both through coastal discharges and through deep water mixing and upwelling [34,35,36]. The SWFS has unique surface currents and a resilient Cross-Shelf Transport Barrier (CSTB) that create a triangular “forbidden zone” isolating the shelf from the GOM loop current [37,38]. This causes suspended sediments, nutrients, dissolved organic matter (DOM) and other contaminants to linger and then accumulate near the coastline rather than being swept away, and may make it especially vulnerable to tropical cyclone disturbances. Despite the well-documented effects of chronic eutrophication and storm pulse events on water quality and HABs in Florida [31,34,35,36,39,40], there has been relatively little investigation into their effects on the benthic communities of the SWFS.

On 28 September 2022, Category 5 Hurricane Ian made landfall along the Southwest Florida coast, near Cayo Costa (26.6190° N, 82.2257° W), with peak winds of 259 km/h and peak storm surge of 4.5 m above ground level [41]. Hurricane Ian was among the most devastating storms in U.S. history, killing over 150 people and causing over $112 billion in damage [41]. Hurricane Ian struck during a survey of water quality and benthic communities of the SWFS (this study, funded by US EPA), providing unique insight into how hurricanes can alter this shelf environment. This is the first study of its kind for the region and one of the first pre-post analyses of tropical cyclone disturbances on benthic communities in shallow continental shelf environments other than coral reefs or estuaries [42,43]. Our objectives for this analysis were to determine the Hurricane Ian’s acute impacts on: (1) substrate characteristics, (2) epibenthic communities, (3) microbial communities, and (4) water quality.

2. Materials and Methods

2.1. Sampling Locations

Eight sites were selected off the coast of Lee and Collier Counties, Florida, USA for assessing the effects of depth and distance from the Caloosahatchee River Estuary (CRE) on water quality, microbial community composition, and benthic habitat characteristics (Figure 1). The sites were arranged along two inshore-offshore transects, running approximately parallel to each other in a west-southwest direction. The northern transect originated near the mouth of the CRE while the south transect originated near Doctors Pass, ~30 km south of the CRE (Figure 1). Transects included sites at the 5, 10, 15, and 20 m depth contours, thus sites were coded by transect identity (N, S) and depth (

Table A1. Depths and locations of each site, and 2022 sampling dates (month/day) for video, water quality, and sediment and microbial community samples. Pre- and Post- indicate pre-Hurricane Ian and post-Hurricane Ian sampling periods, respectively. Blanks indicate where no samples of that type were obtained.

Site	Depth (m)	Latitude and Longitude	Video		Water Quality		Sediment Grain Size		Sediment Microbial		Water Microbial
			Pre-	Post-	Pre-	Post-	Pre-	Post-	Pre-	Post-	Pre-	Post-
N5	3	26.4722° N, 081.9771° W	9/2	11/2	8/23	11/2	9/2		9/2		8/23	11/2
N10	10	26.3086° N, 082.0963° W	9/2	10/21	8/24	10/21	9/2	10/21	9/2	10/21	8/24	10/21
N15	15	26.25751° N, 082.2623° W	9/6	10/22	8/25	10/22	9/6	10/22	9/6	10/22	8/25	10/22
N20	20	26.21683° N, 082.3791° W	9/6	10/22	8/26	10/22	9/6	10/22	9/6	10/22	8/26	10/22
S5	5	26.17203° N, 081.8234° W	9/1	11/2	8/25	11/2	9/1	11/2		11/2	8/25	11/2
S10	10	26.14049° N, 081.9140° W	9/1	10/23	8/26	10/23	9/1	10/23		10/23	8/26	10/23
S15	15	26.05257° N, 082.0984° W	9/8	10/23	8/27	10/23	9/8	10/23		10/23	8/27	10/23
S20	20	25.97247° N, 082.2164° W	9/8	10/23	8/28	10/23	9/8	10/23		10/23	8/28	10/23

, Figure 1). Where possible, hard substrate sites were selected at each depth, however, both 15 m sites and the southern 20 m site were soft substrate. The N5 site was located at 3 m depth due to there being no suitable hard bottom sites at 5 m depth near the CRE mouth. This analysis focuses on the acute effects of Hurricane Ian and is therefore restricted to the sampling events immediately pre- and post-hurricane; August 2022 and October 2022 (Table A1). The pre-hurricane surveys were conducted with small boats operated by Florida Gulf Coast University (FGCU) and the post-hurricane sampling was conducted onboard the R/V Hogarth operated by the Florida Institute of Oceanography (FIO).

2.2. Water Quality Parameters

At each site visit, vertical water column profiles for salinity (PSU), temperature (°C), dissolved oxygen (DO)(mg/L) and turbidity (NTU) were measured using an EXO II multiparameter water quality sonde (SKU #:599502-00, YSI, Yellow Springs, OH, USA). Discrete grab samples of water were collected from the surface (~0.5 m deep), middle depth (half of the total depth at the site), and near bottom, using a Van Dorn water sampler (item #:78902, Forestry Suppliers, Jackson, MS, USA), or with Niskin bottles, and stored following US Environmental Protection Agency (EPA) protocols until lab analyses for chlorophyll a, chromophoric dissolved organic matter (CDOM), silica (SiO₂), nitrate (NO₃⁻), nitrite (NO₂⁻), ammonia (NH₃), orthophosphate (PO₄³⁻), total nitrogen (TN), and total phosphorus (TP). The water samples for chlorophyll a were extracted using EPA method 445.0, and analyzed using a Trilogy fluorometer (model #7200-002, Turner Designs, San Jose, CA, USA) with a chlorophyll acid module (Turner Designs, model #7200-040). CDOM was measured with a Trilogy fluorometer (Turner Designs, model #7200-002) using seawater that was filtered using a 0.45 µm nylon syringe filter and measured with a CDOM module (Turner Designs, model #7200-069-W). Water samples for dissolved nutrients (SiO₂, NO₃⁻, NO₂⁻, NH₃, PO₄³⁻) were filtered using a 0.45 µm nylon syringe filter, put on ice, and frozen until analyzed using a SEAL AA500 continuous flow autoanalyzer (SEAL Analytical Inc., Mequon, WI, USA) following the manufacturer’s methods and calibrations (A-006-19, A-044-19, A-043-19, and A-005-19, respectively). Unfiltered water samples were collected and stored frozen for the analysis of TN and TP, respectively using a Shimadzu TOC with TNM-L accessory (Shimadzu Scientific Instruments, Inc., Columbia, MD, USA) and a SEAL AA500 continuous flow autoanalyzer (A-005-19). In addition to sonde casts and water sampling at each site, lux (lumens/m²), and current velocity were logged continuously throughout the study period at one site (N20, Figure 1). Light data were recorded with an Onset© HOBO UA-002-64 Pendant (HOBO Data Loggers, Bourne, MA, USA), and current data were recorded with a Lowell Instruments TCM-1 tilt current meter (Lowell Instruments LLC, East Falmouth, MA, USA). Total cell counts (TCC) were measured using a BD Accuri C6 flow cytometer (BD Biosciences, Franklin Lakes, NJ, USA). Bacteria TCC was measured by staining with SYBR green, and picocyanobacteria counts were measured through autofluorescence using cell size constraints to target prokaryotes.

2.3. Benthic Video Surveys

Benthic surveying was performed by divers with an Olympus^® TG-6 underwater camera with PT-059 housing (Olympus Corporation of the Americas, Breinigsville, PA, USA) and two Backscatter^® 1000 lumen mounted lights (Backscatter Underwater Video & Photo, Monterrey, CA, USA). Divers took video surveys along two 50-m transects; one at a compass heading of 0° from a cinderblock marking the center of the site, and the other at 270° degrees from the block. Video data were downloaded and analyzed in the laboratory to estimate percent cover by substrate type (hard bottom, shell/hash, and sand/mud) and by seven biotic cover categories (stony coral, soft coral, sponge, sessile invertebrates, seagrass, calcareous green algae [CGA], and fleshy macroalgae). These estimates were made for each 5 m segment along each of the two 50 m transects. Segments were treated as independent statistical replicates (n = 20 per site per survey).

2.4. Sediment Characteristics

Surface sediments were collected in 50 mL vials, either directly by divers or by sub-sampling from a 860-A10 Shipek^® grab sampler (WildCo^®, Yulee, FL, USA), onboard R/V Hogarth, and frozen until further analysis. Due to logistical challenges, post-hurricane sediment samples were not obtained from site N5. Sediments were analyzed for moisture content, bulk dry density, bulk organic matter content via loss of ignition (LOI) and grain size following previously described methods [45,46,47]. For organic matter characterization, at least two grams of dried sediments were combusted at 360 °C for three hours. This lower temperature was employed to minimize the contribution of carbonate loss to the organic matter calculation [47]. All moisture content and bulk organic matter analyses were performed in triplicate. To determine grain size distribution at each site, approximately 10 g of sediment dried at 105 °C was sieved into the following fractions; shell (>4 mm), fine shell (4–2 mm), coarse sand (2–0.25 mm), fine sand (250–63 µm), and silt/clay (<63 µm). Each size fraction was weighed and divided by the total sample weight to get its percent contribution to the sample.

2.5. Microbial Community Analysis

Sediment microbial community analyses were performed with sub-samples of the surface sediments collected, which were frozen at −20 °C. No sediment microbial data were obtained from south transect sites pre-hurricane (August 2022) because those samples dried out in storage prior to analysis, and no sediment microbial data were obtained from site N5 post-hurricane (October 2022) because no sediment sample was taken at that site (Table A1). Water column microbial community data was obtained from all sites. Surface water (200 mL) was filtered through cellulose nitrate filters (0.22 µm, 47 mm diameter, Thermo Scientific™ Nalgene™ analytical test filter funnels) (Thermo Fisher Scientific, Waltham, MA, USA) to collect microbial cells, and filters were frozen at −20 °C. For both sediment and water column microbial samples, amplification of the V4–V5 region of the 16S rRNA gene was conducted using the primer pair 515yF and 926pfR [48]. High-throughput DNA sequences were determined using Illumina MiSeq (2 × 300 bp) (MR DNA, Shallowater, TX, USA). A 30-cycle PCR (HotStarTaq Plus Master Mix Kit, Qiagen, MD, USA) was performed under the following conditions: 95 °C for 5 min, followed by 30 cycles of 95 °C for 30 s, 53 °C for 40 s, and 72 °C for 1 min, after which a final elongation step at 72 °C for 10 min was performed. After amplification, PCR products were checked in 2% agarose gel. Samples were multiplexed using unique dual indices and pooled together in equal proportions based on their molecular weight and concentrations of DNA. Pooled samples were purified using calibrated AMPure XP beads (Beckman Coulter Life Sciences, Brea, CA, USA). Then the pooled and purified PCR product was used to prepare an Illumina DNA library. Demultiplexed FASTQ files were imported to Qiime2 v2023.05 [49] and primer sequences were removed with q2-cutadapt. Quality control and denoising was done with Dada2 [50] to generate amplicon sequencing variants (ASV). Naïve Bayes classification was performed with the q2-feature-classifier plugin using the Silva database (v138.1) for reference [51]. Reference Sequence Annotation and Curation Pipeline (RESCRIPt) [52] was used to optimize classification to the 515f/926r primer set. ASV tables and taxonomy tables were exported and later used for statistical analysis in R and R studio (version 4.2.2). The amplicon sequence datasets were deposited in GenBank under BioProject number PRJNA1043673.

2.6. Statistical Analysis

All statistical analysis was conducted in R (version 4.2.2) [53]. Univariate techniques were used for response variables from water quality, sediment, and benthic video analyses, while multivariate techniques were used for microbial community analyses. The effect of date (pre- vs. post-hurricane), transect (N vs. S), and their interaction was examined for all response variables, except for microbial alpha diversity, for which only date was considered (

Table A3. Statistical analyses applied, and results of analyses, for each response variable from benthic video transects. (*) indicates p-value < 0.05, (**) indicates p-value < 0.01, (***) indicates p-value < 0.001. Surface water (SW) and surface sediment (Sed) abbreviated in the response variable column and south transect (ST) is abbreviated in the samples excluded column. A two-way permutational ANOVA has the same interpretation of a parametric two-way ANOVA but the p-value is calculated differently. The observed data is used to calculate the f-value as in a traditional two-way ANOVA, then the response variable is repeatedly shuffled randomly (i.e., iterations) and assigned to different sample-IDs while keeping group sizes constant. For each iterated data set, a f-value is calculated and compared to the observed f-value. The p-value is the proportion of iterations that have a test statistic equal to or larger than that of the observed data.

Response Variable	Factors & Interactions	Sample Size (n)	Test	p-Values	Samples Excluded
Benthos and Substrate
Hard Bottom Cover	Date Transect Date × Transect	300	Permutation Two-way ANOVA	<2 × 10−16 *** <2 × 10−16 *** 0.5811	N/A
Total Biotic Cover	Date Transect Date × Transect	300	Permutation Two-way ANOVA	<2 × 10−16 *** <2 × 10−16 *** 0.6545	N/A
Stony Coral % Cover	Date Transect Date × Transect	300	Permutation Two-way ANOVA	<2 × 10−16 *** 0.3194 0.1061	N/A
Soft Coral % Cover	Date Transect Date × Transect	300	Permutation Two-way ANOVA	0.8235 <2 × 10−16 *** 0.6545	N/A
Sponge % Cover	Date Transect Date × Transect	300	Permutation Two-way ANOVA	0.00620 ** 0.3402 0.1714	N/A
Sessile Invertebrate % Cover	Date Transect Date × Transect	300	Permutation Two-way ANOVA	<2 × 10−16 *** 0.1401 0.1510	N/A
Seagrass % Cover	Date Transect Date × Transect	300	Permutation Two-way ANOVA	0.0008 *** 0.0040 ** 0.0112 *	N/A
Calcareous Green Algae % Cover	Date Transect Date × Transect	300	Permutation Two-way ANOVA	<2 × 10−16 *** 0.0024 ** 0.0102 *	N/A
Fleshy Macroalgae % Cover	Date Transect Date × Transect	300	Permutation Two-way ANOVA	<2 × 10−16 *** <2 × 10−16 *** 0.2377	N/A
Water Quality and Nutrients
Chlorophyll a	Date Transect Date × Transect	48	Permutation Two-way ANOVA	0.9804 0.8627 0.6429	N/A
Ammonia (NH3)	Date Transect Date × Transect	48	Permutation Two-way ANOVA	0.0044 * 0.4622 0.4444	N/A
Nitrate (NO3)	Date Transect Date × Transect	48	Permutation Two-way ANOVA	0.2922 0.2182 0.7647	N/A
Nitrite (NO2)	Date Transect Date × Transect	48	Permutation Two-way ANOVA	<0.001 ** 0.3924 0.1523	N/A
Orthophosphate (oP)	Date Transect Date × Transect	48	Permutation Two-way ANOVA	0.0146 * 0.2153 0.7451	N/A
Silica (SiO2)	Date Transect Date × Transect	48	Permutation Two-way ANOVA	0.0158 * 0.0548 0.0633	N/A
Total Nitrogen (TN)	Date Transect Date × Transect	48	Permutation Two-way ANOVA	<0.0001 *** 0.6667 0.9804	N/A
Total Phosphorus (TP)	Date Transect Date × Transect	48	Permutation Two-way ANOVA	<0.001 ** <0.001 ** <0.001 **	N/A
CDOM (PPB)	Date Transect Date × Transect	48	Permutation Two-way ANOVA	<0.001 ** 0.8824 0.5542	N/A
Turbidity (NTU)	Date Transect Date × Transect	48	Permutation Two-way ANOVA	<0.001 ** 0.1924 0.8039	N/A
Sediment Characteristics
Volumetric Water Content (%)	Date Transect Date × Transect	7	Permutation Two-way ANOVA	<0.001 * 0.3552 0.9608	N5
Organic Matter Content (%)	Date Transect Date × Transect	7	Two-way ANOVA	0.0124 * 0.1515 0.9623	N5
Shell fraction (%)	Date Transect Date × Transect	7	Permutation Two-way ANOVA	0.8039 1.0000 0.2170	N5
Fine shell fraction (%)	Date Transect Date × Transect	7	Permutation Two-way ANOVA	0.6429 0.5412 0.8431	N5
Coarse sand fraction (%)	Date Transect Date × Transect	7	Permutation Two-way ANOVA	0.3043 0.5412 0.7255	N5
Fine sand fraction (%)	Date Transect Date × Transect	7	Permutation Two-way ANOVA	0.9216 0.0188 * 0.4906	N5
Silt fraction (%)	Date Transect Date × Transect	7	Permutation Two-way ANOVA	0.0570 0.0140 * 0.1230	N5
Microbial Alpha Diversity and Abundances
Surface water -Richness	Date	8	Paired T-test	<0.001 **	N/A
Surface water -Hill-Shannon	Date	8	Paired T-test	<0.001 **	N/A
Surface water -Hill-InvSimpson	Date	8	Paired T-test	0.011 *	N/A
Sediment-Richness	Date	6	Paired T-test	0.095	N5, ST
Sediment -Hill-Shannon	Date	6	Paired T-test	0.187	N5, ST
Sediment -Hill-InvSimp	Date	6	Paired T-test	0.717	N5, ST
Pre-SW + Sed Richness	Sample type	8	Paired T-test	<0.001 **	ST
Pre-SW + Sed Hill-Shannon	Sample type	8	Paired T-test	<0.001 **	ST
Pre-SW + Sed Hill- InvSimp	Sample type	8	Paired T-test	<0.001 **	ST
Post-SW + Sed Richness	Sample type	14	Two-Sample Fisher-Pitman Permutation Test	0.5664	N5
Post-SW + Sed Hill-Shannon	Sample type	14	Two-Sample Fisher-Pitman Permutation Test	0.2805	N5
Post-SW + Sed Hill-InvSimp	Sample type	14	Two-Sample Fisher-Pitman Permutation Test	0.1806	N5
Total cell count (surface water)	Date Transect Date × Transect	16	2-way ANOVA	0.231 0.295 0.848	N/A
Picocyanobacteria counts (surface water)	Date Transect Date × Transect	16	2-way ANOVA	0.0034 * 0.5528 0.4950	N/A

). For water quality parameters collected at surface, middle, and bottom depths, the mean of all depths per site, per date was used as the response value. When the data did not meet assumptions of parametric statistics (i.e., normality and homogeneity of variance), non-parametric, permutational alternatives were used. For two-way ANOVA, the R package LmPerm [54] was utilized (Table A3). To allow even comparisons of sediment characteristics and sediment alpha diversity between pre- and post-hurricane conditions, statistical analysis was conducted only on sites with both pre- and post-hurricane data. A paired t-test was performed on alpha diversities to compare pre- and post-hurricane microbial communities, and if assumptions were not met the permutational alternative was used from the R package coin [55].

Microbial multivariate statistics were performed on the ASV level. Data manipulation and management were done using the phyloseq package [56]. Hill alpha diversities were computed with the HillR package [57] with samples rarefied to the lowest sequencing depth for equal coverage, statistical analyses were conducted only on sites with samples for both pre- and post-hurricane data (

Table A4. Hill Diversity values for surface water microbial communities pre- and post-hurricane (August 2022 “Pre” and October 2022 “Post”). Mean values across all samples pre- and post- hurricane shown with standard deviation.

	Species Richness		Hill-Shannon		Hill-InvSimpson
Site	Pre	Post	Pre	Post	Pre	Post
N5	287	353	94.15	109.23	44.17	45.66
N10	263	386	66.75	127.61	28.36	61.34
N15	287	448	103.12	180.82	56.63	100.55
N20	247	373	82.15	146.67	42.48	72.20
S5	226	486	53.31	118.35	16.89	38.64
S10	240	389	55.87	109.94	18.85	41.37
S15	253	347	93.54	133.49	49.54	77.27
S20	272	369	87.82	144.27	42.96	83.38
Mean	259 ± 22	394 ± 48	79.59 ± 18.72	133.80 ± 23.71	37.48 ± 14.47	65.05 ± 22.18

and

Table A5. Hill Diversity values for surface sediment microbial communities pre- and post-hurricane (August 2022 “Pre” and October 2022 “Post). Mean and standard deviation are shown for just N10, N15, and N20, as these were the only sites for which sediment microbial community data were obtained both before and after the hurricane. Not determined abbreviated as n.d.

	Species Richness		Hill-Shannon		Hill-InvSimpson
Site	Pre	Post	Pre	Post	Pre	Post
N5	1606	n.d	909.79	n.d	411.27	n.d
N10	1098	1213	666.69	867.73	337.06	589.46
N15	1339	371	914.63	261.69	585.9	174.09
N20	1799	167	813.4	95.87	135.6	51.07
S5	n.d	994	n.d	470.06	n.d	92.65
S10	n.d	673	n.d	426.79	n.d	239.42
S15	n.d	461	n.d	302.1	n.d	159.46
S20	n.d	310	n.d	105.42	n.d	15.78
Mean of N10, N15, N20	1412 ± 356	584 ± 554	798.24 ± 124.66	408.45 ± 406.31	352.85 ± 225.56	271.54 ± 282.11

). The vegan package was used to calculate Bray-Curtis dissimilarity from relative abundances, which was the beta diversity metric utilized in non-metric multidimensional scaling (NMDS) [58]. Within vegan, the function adonis2 was used for permutational multivariate analysis of variance (PERMANOVA), to test if centroids of tested groups are statistically different from each other. If the samples did not display equal dispersion (assessed with betadisper) between groups PERMANOVA was not conducted. All graphics were made with ggplot2 [59].

3. Results

3.1. Physical Effects of Hurricane Ian

On 28 September 2022, the eyewall of Hurricane Ian came within 12 km of the N20 site (Figure 1). The tilt current meter deployed at the site measured a peak value of benthic current velocity of 0.69 m/s, far exceeding the previous 10-day mean of 0.037 m/s. Furthermore, the Sanibel Captiva Conservation Foundation (SCCF)’s TRIAXYS© wave buoy, located 31 km NE from N20, recorded 7.2 m wave heights and surface current velocity of 2.4 m/s (the highest ever recorded by a TRIAXYS© wave buoy) during the hurricane before breaking loose. The average benthic light value at N20 on the day prior to the hurricane (27 September 2022) was 40.2 lumens/m² compared with an average of 0 lumens/m² on the day of and 9 days following the hurricane. Divers reported near-darkness at depth even at the 10 m sites three weeks after the storm. Other indicators of the severity of the physical disturbance to the benthic environment by Hurricane Ian’s waves include the toppling and >100 m movement of segments of the “Arc Tower” artificial reef; a heavily weighted steel and concrete structure 21 km N from N20 in 18 m of water, which had been in place since 1994.

3.2. Water Quality Parameters

The majority of the water quality parameters measured (salinity, DO, temperature, turbidity, chlorophyll a, CDOM, silica, nitrate, nitrite, ammonia, total nitrogen, total phosphorus, and orthophosphate) changed significantly after Hurricane Ian (Figure 2, Figure 3, Figure A1, Figure A2 and Figure A3). Salinity dropped by 1–2 PSU, and temperature dropped by ~5 °C at all sampling sites (Figure A1 and Figure A2). There was an increase in DO at 10–20 m depths following the hurricane, but a decrease in DO at 5 m depths near shore (Figure 2). Prior to the hurricane there was strong stratification at sites N20 and N15 with near hypoxic conditions (<3 mg/L O₂) at site N20. This stratification and bottom hypoxia was not observed following the hurricane except at S5 (Figure 2).

Post-hurricane turbidity was higher than pre-hurricane values for all sampling sites, with the exception of N5, where turbidity was elevated both pre- and post-hurricane (Figure 3). The overall increase in turbidity was significant (df = 1, 44, Iter = 5000; p < 0.001). Chlorophyll a was higher closer to shore both pre- and post-hurricane (Figure 3 and Figure A3). There was a decrease in chlorophyll a along the north transect after the hurricane, and an increase across the south transect, but the difference was not statistically significant (Table A3). Prior to the hurricane CDOM was slightly elevated at nearshore sites. After the hurricane CDOM increased significantly (df = 1, 44, Iter = 5000; p < 0.001), although the increase was not observed at sites N5 and S5. Post-hurricane sampling at N5 and S5 was performed two weeks after sampling other sites because they were too shallow to safely sample from the R/V Hogarth. Thus, the lower CDOM values for these inshore sites could be attributed to the difference in sampling period allowing extended dilution, photochemical degradation, and/or salting out of the CDOM over those two weeks (Figure 3 and Figure A3).

Silica concentrations significantly decreased across all sites post-hurricane, except for the nearshore site N5 where concentrations increased across all depths (df = 1, 44; Iter = 5000; p = 0.0158). Overall, nitrate concentrations increased across most sites post-hurricane, but these changes were not significant (Figure 3 and Figure A3, Table A3). Similarly, nitrite concentrations increased across all stations post-hurricane with the highest increases at the offshore (10–20 m) north transect sites, and the nearshore south transect site (S5). These nitrite changes were significant (df = 1, 44; Iter = 5000; p < 0.001). Ammonia concentrations increased across most sites except for site N20 and S20 where there were no major changes in concentrations observed. Overall, ammonia concentrations across all sites were significantly higher post- hurricane (df = 1, 44; Iter = 5000; p = 0.0044). Orthophosphate concentrations increased at sites N5, N15, N20, and S5 and the pre- and post-hurricane change was significant (df = 1, 44; Iter = 5000; p = 0.0146). Total nitrogen concentrations decreased across all sites post-hurricane (df = 1, 44; Iter = 5000; p < 0.001). Total phosphorus concentrations also decreased across all sites post-hurricane, with a greater drop in TP along the north transect than the south transect. The differences in TP concentrations between transect and date were significant and there was a significant transect x date interaction (df = 1, 44; Iter = 5000; p < 0.001) (Figure 3, Table A3).

3.3. Epibenthos and Substrate

There was a significant decrease in hard bottom cover after the passage of Hurricane Ian (df = 1, 296; Iter = 5000; p < 0.001), and there was visual evidence of burial and scour (Figure 4a, Table A3). Site N20 lost the most hard bottom habitat post-hurricane while N5 gained the most (Figure 4a). Total biotic cover declined significantly after the hurricane (df = 1, 296; Iter = 5000; p < 0.001) (Figure 4b, Table A3). The highest biotic cover both pre- and post-hurricane was at N20 and S5 (21.5% and 54.5%, respectively, post-hurricane) (Figure 4b). The greatest relative decrease in biotic cover was at S10 and N10 (98.3% and 92.6% decrease, respectively) such that remaining biotic cover was near zero at these sites, 0.5% and 1.7%, respectively (Figure 4b).

Sponge, stony coral, sessile invertebrates, seagrass, CGA, and fleshy macroalgae significantly decreased after the hurricane (45%, 71%, 77%, 100%, 100%, and 58% loss, respectively) while changes in soft coral were not significant by date (17%) (Figure 5a,b, Table A3). Seagrass (Halophila decipiens) was observed at N5 pre-hurricane, but was absent from all sites post-hurricane (Figure 5a,b). Fleshy macroalgae, the dominant biotic cover category at many sites pre- and post-hurricane, was reduced at all sites post-hurricane (df = 1, 296; Iter = 5000; p < 0.001) (Table A3).

3.4. Sediment Characteristics

The median silt fraction was 2.64% (IQR = 3.1) before the hurricane and 5.79% after (20.2), but this overall date effect was marginally non-significant (df = 1,10; Iter = 1658; p = 0.057) (Table A3, Figure A5). The north transect had higher silt content (median (IQR) = 10.33% (31.1)) than the south transect (2.64% (4.58) (df = 1,10; Iter = 5000; p = 0.014)). There was no significant interaction of date and transect on silt fraction (df = 1,10; Iter = 715; p = 0.123). There was no region-wide impact of the hurricane on the other sediment size fractions (Table A3), although compositional changes did occur at the scale of individual sites (Figure A4a,b). Pre-hurricane, sediment organic matter content (mean ± SD = 1.78 ± 0.57%) was higher than post-hurricane (0.99 ± 0.40%) (df = 1,10; F = 9.259; p = 0.012). The north and south transects had similar organic matter (df = 1,10; F = 2.41; p = 0.151) and there was no differential impact of the hurricane on organic matter between the two transects (df = 1,10; F = 0.002; p = 0.96) (Figure A5). Post-hurricane, sediment moisture decreased (df = 1,10; Iter = 5000; p < 0.001). There was no difference in sediment moisture between the north and south transects (df = 1,10; Iter = 183; p = 0.355), and there was no differential impact within the transects due to the hurricane (df = 1,10; Iter = 51; p = 0.961) (Figure A6).

3.5. Microbial Community Structure

Prior to the hurricane, microbial alpha diversity was lower in the surface water than in the surface sediment (Hill-Shannon: t = 12.63, df = 6, p < 0.001. Richness: t = 7.74, df = 6, p < 0.001. Hill-InvSimp: t = 3.47, df = 6, p < 0.001) (Table A4 and Table A5). Post-hurricane, this difference in diversity between the sediment and water column was not observed (Hill-Shannon: t = 1.16, Iter = 9999, p = 0.281. Richness: t = 0.603, Iter = 9999, p = 0.566. Hill-InvSimp: t = 1.368, Iter = 9999, p = 0.181). Water column microbial diversity increased post-hurricane at all sites (Hill-Shannon: t = 5.10, df = 14, p < 0.001. Richness t = 7.15, df = 14, p < 0.001. Hill-InvSimp: t = 2.94, df = 14, p = 0.011) (Table A4). Mean sediment microbial diversity decreased post-hurricane (Table A5), but this difference was not significant (Hill-Shannon: t = 1.59, df = 4, p = 0.18. Richness: t = 2.18, df = 4, p = 0.095. Hill-InvSimp: t = 0.39, df = 4, p = 0.717). Microbial abundances were also determined in the surface waters (Figure A8). For total cell counts there was no difference were found between pre- or post- hurricane (df = 1,12; F = 1.594; p = 0.231) or between transects (df = 1,12; F = 1.20; p = 0.295). Picocyanobacteria counts were lower post-hurricane (df = 1,12; F = 13.04; p = 0.004) with no difference between transects (df = 1,12; F = 0.37; p = 0.553).

Microbial community composition also differed before and after the hurricane. Pre-hurricane surface water microbial communities were widely spread in ordination space, with a pattern reflecting their distance from shore and the mouth of the CRE. Site N5, near the mouth of the CRE, had the most distinct surface water microbial community. Post-hurricane surface water microbial communities were more clustered in ordination space and differed significantly from pre-hurricane communities (PERMANOVA, R² = 0.267, p = 0.001), but still exhibited some inshore-offshore differentiation (Figure 6a). Comparisons among sediment microbial communities were complicated by the lack of pre-hurricane data from the south transect, and the lack of post-hurricane data for site N5. However, among sites sampled on both dates, there was a notable shift in sediment microbial communities and greater differentiation among sites post-hurricane (Figure 6b and Figure 7,

Table A7. Relative abundance (%) of the top 12 taxa at the class level in surface sediment microbial communities in the pre-hurricane (August 2022, “Pre”) and post-hurricane (October 2022, “Post”) samples. Mean for just N10, N15, and N20 is shown, as these were the only sites for which sediment microbial community data were obtained both before and after the hurricane. Not determined abbreviated as n.d. Mean values shown with standard deviation.

Taxon	Time	N5	N10	N15	N20	S5	S10	S15	S20	Mean N10, N15, N20
Gammaproteobacteria	Pre	18.23	17.36	13.95	12.82	n.d.	n.d.	n.d.	n.d.	14.71 ± 2.36
	Post	n.d.	11.22	15.46	12.58	20.40	12.37	19.20	38.49	13.09 ± 2.16
Planctomycetes	Pre	5.56	9.43	14.78	15.11	n.d.	n.d.	n.d.	n.d.	13.11 ± 3.19
	Post	n.d.	5.85	8.40	2.75	10.40	7.99	8.50	6.64	5.67 ± 2.83
Alphaproteobacteria	Pre	14.75	12.60	9.32	17.13	n.d.	n.d.	n.d.	n.d.	13.02 ± 3.92
	Post	n.d.	6.97	9.45	6.74	10.81	7.07	10.52	4.23	7.72 ± 1.51
Bacteroidia	Pre	16.11	10.10	11.94	11.16	n.d.	n.d.	n.d.	n.d.	11.07 ± 0.92
	Post	n.d.	11.85	9.72	10.23	12.41	7.09	6.82	1.74	10.60 ± 1.11
Acidimicrobiia	Pre	3.29	6.08	3.78	3.24	n.d.	n.d.	n.d.	n.d.	4.36 ± 1.51
	Post	n.d.	2.16	5.08	3.24	5.70	4.67	5.91	3.59	3.49 ± 1.48
Anaerolineae	Pre	4.35	4.42	3.82	3.73	n.d.	n.d.	n.d.	n.d.	3.99 ± 0.37
	Post	n.d.	5.67	4.00	5.73	4.38	10.72	6.01	3.42	5.13 ± 0.98
Thermoanaerobaculia	Pre	2.73	2.85	4.61	3.24	n.d.	n.d.	n.d.	n.d.	3.57 ± 0.92
	Post	n.d	2.55	4.93	4.03	2.89	4.89	4.28	7.63	3.84 ± 1.20
Desulfobacteria	Pre	5.49	3.11	4.40	4.42	n.d.	n.d.	n.d.	n.d.	3.98 ± 0.75
	Post	n.d.	4.19	2.02	5.80	1.28	6.44	2.51	2.64	4.00 ± 1.90
Bacilli	Pre	1.00	2.13	1.54	1.38	n.d.	n.d.	n.d.	n.d.	1.68 ± 0.40
	Post	n.d.	3.27	5.57	11.75	1.49	3.91	4.56	4.81	6.86 ± 4.39
Actinobacteria	Pre	1.44	3.07	0.53	0.53	n.d.	n.d.	n.d.	n.d.	1.38 ± 1.47
	Post	n.d.	1.28	4.67	12.44	1.87	1.56	2.31	3.75	6.13 ± 5.72
Clostridia	Pre	0.66	2.53	0.91	0.86	n.d.	n.d.	n.d.	n.d.	1.43 ± 0.95
	Post	n.d.	3.74	1.95	0.12	2.33	2.02	1.80	1.66	1.93 ± 1.81
Desulfobulbia	Pre	2.70	3.12	2.98	3.79	n.d.	n.d.	n.d.	n.d.	3.30 ± 0.43
	Post	n.d.	2.78	2.43	1.07	1.09	2.53	1.95	0.31	2.09 ± 0.90

). Sites N15 and N20, which had similar sediment microbial communities pre-hurricane had divergent communities post-hurricane (Figure 6a).

In the surface sediments, the dominant contributors to the microbial community were Gammaproteobacteria, Planctomycetes, Alphaproteobacteria, and Bacteroidia, but these taxa were impacted by the hurricane. Gammaproteobacteria in the sediments increased post hurricane, from an average of 15.6% to 18.5% relative abundance (Figure 7b, Table A7). Planctomycetes (11.2–7.2%), Alphaproteobacteria (13.8–7.9%), and Bacteroidia (12.3–8.6%) all decreased in relative abundance post-hurricane. The minor contributors that increased in relative abundance after the hurricane were Bacilli (1.5–5.0%), Actinobacteria (1.4–4.0%) and Clostridia (1.2–1.9%). The increases in relative abundance by Gammaproteobacteria and the minor contributors were largely driven by offshore sites, in particular N20 and S20 (Figure 7b, Table A7), but a fair pre-post hurricane comparison of these taxa can only be made for sites N10, N15, and N20 due to missing samples from the other sites (Table A7).

The surface water microbial communities were dominated by Alphaproteobacteria, Bacteroidia, Cyanophyceae, and Gammaproteobacteria, but relative abundances of these groups shifted after the hurricane (Figure 7a,

Table A6. Relative abundance (%) of the top 10 taxa at the class level in surface water microbial communities in the pre-hurricane (August 2022, “Pre”) and post-hurricane (October 2022, “Post”) samples. Mean values shown with standard deviation.

Taxon	Time	N5	N10	N15	N20	S5	S10	S15	S20	Mean
Alphaproteobacteria	Pre	38.48	26.42	41.72	38.57	28.74	28.95	41.60	42.17	35.83 ± 6.64
	Post	29.73	34.68	31.77	34.45	25.26	32.43	39.01	30.73	32.26 ± 4.04
Cyanophyceae	Pre	3.93	23.01	13.36	21.66	24.05	22.73	17.02	15.89	17.70 ± 6.78
	Post	17.63	12.44	6.78	12.50	22.43	19.81	2.41	6.14	12.52 ± 7.11
Bacteroidia	Pre	31.73	25.92	16.22	13.66	17.43	17.86	18.36	11.62	19.10 ± 6.60
	Post	18.02	23.04	22.21	15.76	24.74	18.10	24.92	25.24	21.50 ± 3.70
Gammaproteobacteria	Pre	12.42	4.46	4.82	4.45	5.39	5.97	4.71	7.18	6.18 ± 2.69
	Post	19.63	11.45	14.57	10.35	9.06	8.71	15.10	13.38	12.78 ± 3.66
Verrucomicrobiae	Pre	2.00	0.48	6.30	8.88	0.59	0.53	2.69	7.56	3.63 ± 3.43
	Post	0.45	1.55	2.43	1.35	0.18	1.11	3.36	4.65	1.88 ± 1.52
Acidimicrobiia	Pre	2.90	3.72	3.23	2.72	3.61	3.70	2.18	4.16	3.28 ± 0.65
	Post	3.47	4.44	3.01	3.42	2.17	3.59	1.37	3.42	3.11 ± 0.94
Planctomycetes	Pre	1.52	2.83	1.24	1.37	2.59	2.89	2.44	0.35	1.90 ± 0.92
	Post	1.65	3.51	5.32	6.44	5.17	2.98	2.72	2.35	3.77 ± 1.68
Thermoplasmata	Pre	1.53	1.34	0.52	0.03	1.02	0.48	0.94	1.23	0.89 ± 0.51
	Post	2.83	3.31	4.51	7.17	0.28	2.56	4.99	8.63	4.29 ± 2.67
Rhodothermia	Pre	0.52	6.27	3.06	1.80	5.02	11.12	2.79	2.33	4.11 ± 3.36
	Post	0.52	0.49	0.30	0.12	0.72	0.73	0.23	0.15	0.41 ± 0.24
Marinimicrobia (SAR406_clade)	Pre	0.05	0.55	3.81	2.73	1.45	0.25	2.38	4.44	1.79 ± 1.78
	Post	0.19	0.40	0.61	0.85	1.67	0.68	0.55	0.58	0.51 ± 0.22

). Cyanophyceae had one of the largest changes in relative abundance, declining from an average of 17.7% pre-hurricane to 12.5% post-hurricane. Alphaproteobacteria also declined from 35.8% pre-hurricane to 32.3% post-hurricane. Both Gammaproteobacteria and Bacteroidia increased in relative abundance after the hurricane, 6.2% to 12.8% and 19.1% to 21.5%, respectively. Less abundant taxa in the water column also changed in relative abundance before and after the hurricane, such as Planctomycetes (1.9–3.8%) and the Archaea Thermoplasmata (0.89–4.3%) increased in relative abundance while Verrucomicrobiae (3.6–1.9%) and Rhodothermia (4.1–0.41%) declined.

4. Discussion

4.1. Hurricane Ian in Historic Context

The extent of the environmental damage wrought by tropical cyclones depends on individual storm characteristics like wind intensity, landfall location, and precipitation [5,60]. Hurricane Ian caused a higher surge (4.5 m) and caused more extensive coastal damage than other hurricanes that made landfall in the region in recent decades, despite not possessing the strongest winds or lowest pressures. For example, Hurricane Charley (2004, Category 4 at landfall near Cayo Costa, FL) had a trajectory and intensity similar to Hurricane Ian, but was more compact and faster moving, generating a maximum storm surge of 2.5 m [61]. Hurricane Wilma (2005, Category 3 at landfall near Cape Romano, FL) was weaker, but produced a storm surge of 2.7 m [62]. Hurricane Irma (2017, Category 3 at landfall near Cudjoe Key, FL) crossed just inland of the southwest Florida coastline and was associated with a major freshwater discharge disturbance, but not a significant storm surge or wave disturbance on the SWFS [63]. To summarize, Hurricane Ian’s disturbance to the SWFS was not exceptional in terms of freshwater discharge but was more severe in terms of waves and storm surge than other storms of the prior two decades [64].

4.2. Changes in Water Quality

Images captured by NASA’s MODIS satellite two days after Hurricane Ian’s landfall indicate reduced optical water quality due to both wave action and coastal discharge (Figure A7). Cloudy waters offshore are attributable to the resuspension of carbonate-rich seafloor sediments by deep wave bases, while dark-hued plumes extending from the mouths of estuaries onto the shelf indicate the discharge of freshwater, CDOM, and various pollutants from the Myakka, Peace, and Caloosahatchee Rivers [65]. This pattern was also evident in our observations as increases of both turbidity and CDOM following the hurricane (Figure 3).

Discharges from Lake Okeechobee to the CRE through an artificial connection add to local watershed flows and greatly influence physico-chemical properties of the estuary and SWFS waters [66]. During Hurricane Ian, freshwater discharge from the Franklin Lock and Dam (S-79) water control structure to the CRE surpassed optimal flow levels for the estuary’s health (73.6 m³/s) by eleven-fold, peaking at over 680 m³/s [67,68]. While the most extreme flows were brief, they were followed by a prolonged period of >73.6 m³/s S-79 discharges to the SWFS. The 60-day mean discharge from S-79 on 31 August 2022 (during the pre-hurricane sampling period for this study) was 29.95 m³/s. By contrast it was 111.9 m³/s on October 31st (during the post-hurricane sampling period) [67,68]. Reductions in salinity and increases in CDOM throughout the SWFS study region are likely explained by this large introduction of freshwater.

In addition to reducing salinity, tropical cyclones can rapidly reduce sea surface temperatures by 3–5 °C through deep mixing by large waves and entrainment of upwelled cold waters [69,70,71,72,73]. This, in combination with seasonal cooling, likely explains the 5 °C post-hurricane temperature reduction in this study (Figure A2). The mixing and cooling effect of the hurricane may also have been responsible for the increased DO concentrations and abatement of bottom hypoxia on the SWFS after the storm. This increase contrasts with a decrease in DO typically observed in Southwest Florida estuaries after hurricanes [74]. The difference between shelf and estuarine DO responses might be explained by predominance of physical mixing effects on the shelf versus predominance of increased biological oxygen demand and salinity stratification in estuaries [74,75]. Our nearshore sites (N5 and S5) did exhibit a modest decrease in DO post-hurricane, which is consistent with the notion that DO drops are more inshore, estuarine dynamics. However, because these sites were sampled two weeks later than the other sites it is uncertain whether the drop was site-related or time-related.

The increased concentrations of nutrients (nitrate, nitrite, orthophosphate, and ammonia) following Hurricane Ian (Figure 3 and Figure A3) are consistent with previously reported post-hurricane changes [18,76]. These increases can be attributed to a combination of factors, including the decomposition of terrestrial organic matter [77,78,79], nutrient-rich freshwater discharge [80,81], and deep-mixing and benthic flux on the shelf [82,83,84]. The decrease in DO at nearshore stations (N5 and S5) is consistent with decomposition of terrestrial organic matter as a nutrient source. However, the specific contribution of each source remains unclear. In the marine nitrogen cycle, nitrate, nitrite, and ammonia are interconnected and undergo transformations during the breakdown and re-mineralization of organic matter [85,86,87,88]. Ammonia, resulting from the breakdown of organic matter, can be further oxidized into nitrite and then nitrate through the oxidation of ammonia i.e., nitrification [89,90,91]. The rates of nitrification can be influenced by environmental conditions such as DO, light, temperature, salinity, and turbidity [91,92,93], which likely contributes to the observed elevated levels of nitrite and nitrate. In addition to the remineralization or organic matter, the increase of orthophosphate is likely associated with runoff, sediment resuspension and dissolution, and sedimentary fluxes [94]. Furthermore, changes in nutrient concentrations at offshore sites post-hurricane could be influenced by deepwater mixing and upwelling [95]. In contrast, the general decrease in silica concentrations can be attributed to the biological uptake by diatoms, whose abundances are known to be influenced by storm activity [96,97].

Typically, there is an increase in chlorophyll a concentrations following major tropical cyclones [72,98,99,100,101]. The increase in chlorophyll a after storms, may stem from terrigenous and upwelled nutrient inputs, and offshore advection of coastal waters by wind-driven currents [72,102]. However, no clear pattern could be discerned for pre- and post-hurricane changes in chlorophyll a in the current study (Figure 3). This may be due to already-high chlorophyll a concentrations in the study region at the time of the hurricane; typical for late summer in the northeastern GOM [78,103,104,105], and to countervailing effects of increased nutrients and reduced optical water quality in the aftermath of the storm. Remote sensing approaches examining broader regions of the SWFS might reveal aspects of the post-hurricane phytoplankton community response not captured in this more localized, nearshore study [72,106,107].

The timescale for post-hurricane algal bloom development may also exceed the timescale of sampling for this study. Others have associated high freshwater and nutrient flux to the SWFS following hurricanes with HABs that develop several months to a year after the immediate disturbance [32,33].

4.3. Changes in Substrates and Benthos

Hurricane Ian significantly altered benthic habitats in the study region, reducing the availability of hard bottom substrate and most types of biotic cover at the sites surveyed. These changes likely reflect initial scouring away of epibenthic biota by strong currents and suspended debris, and burial by transported and deposited sediments [22,108,109] (Figure A4a,b). The near complete loss of epibiota and hard bottom substrate from both 10 m depth sites, and their replacement with sand and silt, indicates bulk sediment transport and burial to considerable depth. Prior to the hurricane these sites had ~30 cm tall colonies of the scleractinian corals Solenastrea bournoni and Oculina robusta, which appeared to have been completely buried during the storm. Unconsolidated sediment deposits on the SWFS are spatially heterogeneous, including shoreface-connected sand sheets at <8 m depth, and alternating sand ridges (up to 4 m high) and exposed hard bottom beyond the 8 m isobath [110]. Critical threshold velocities for erosion of fine sand from smooth bottom (17.2 cm/s) [111], were far exceeded even at 20 m depth during Hurricane Ian, as indicated by the tilt meter data from Site N20. I.e., there was an ample supply of unconsolidated sediments in the area, and ample energy to move it during the hurricane, to the detriment of the organisms buried at N10 and S10. Further offshore at N20, the percentage of exposed hard bottom habitat also declined sharply, from 35% to 5%, but it appeared less deeply buried, with many soft corals and sponges protruding through the sediment veneer. The lack of deep burial at Sites N20 and S5 might be due to the relatively high relief of the limestone ledge habitats at those sites, or due to other aspects of their structure and setting. Identifying hard bottom features on the SWFS that have persisted and are likely to persist through tropical cyclones could help managers set priority seabed areas for protection from anthropogenic disturbance.

Organisms remaining after the storm included some tall and strongly attached taxa that may be resistant to scour and burial, such as sponges and soft coral (Figure 5a,b), as well as some resilient taxa like fleshy macroalgae that may have quickly regrown or recolonized at sites where hard substrates persisted through the storm [20,21]. The sponge community in the region surveyed was comprised largely of encrusting and firmly-attached forms, which others have found are resistant to tropical cyclone impacts [112]. Some of the biotic cover groups most negatively impacted by Hurricane Ian (seagrass, calcareous green algae, and fleshy macroalgae) have been identified as vulnerable to physical disturbance in other storm disturbance studies, as well [112,113,114]. However, these relatively fast-growing taxa are also known to fluctuate seasonally in this region [115], thus determining if they have suffered a long-term loss will require additional assessment. In ecosystems with pronounced seasonal fluctuations in environmental conditions, such as temperate estuaries, benthic communities may return to baseline conditions quickly after storm disturbance [75]. However, where seasonal fluctuation is less pronounced and benthic ecosystems are comprised of slower-growing, structure-forming species, severe disturbances can lead to enduring changes in ecosystem state [5].

The north transect of this study was likely more severely impacted by Hurricane Ian than the south transect, as the hurricane’s path came closer to the north study sites (Figure 1) and these sites were closer to freshwater discharges from the CRE. The greater contribution of fine-grained sediment along the north transect versus the south is consistent with an influence of sediments borne by freshwater discharge [116]. However, aerial imagery indicates a large amount of sediment resuspension across the entire SWFS (Figure A8), and a resorting of grain sizes due to differential settling velocity would leave smaller sized sediment fractions on top [108]. Thus, the conspicuous, fine-grained sediment deposits seen after the storm are likely from a combination of riverine export and resorting of existing shelf sediments.

Decreased sediment organic matter post-hurricane, may be due to resuspension, decomposition, and advection of low-density organic matter from the system, or due to an influx of fine sediments with higher contributions from inorganic minerals, relative to the original sediments. Moisture content in the sediments also decreased post hurricane, possibly due to mechanisms similar to those that changed the organic matter content [117]. There is precedent for large scale advective transport of fine sediments during tropical cyclone events in shelf waters. For example, in 2005, between Hurricanes Katrina and Rita, 1.16 × 1015 ± 1.56 × 1014 g of sediment were remobilized along the Louisiana and Texas shelves [118]. Further, immediately following Typhoon Chan-hom’s passage in the East China Sea, the suspended sediment concentration increased 50 times compared to pre-typhoon concentrations and was transported both along-shelf and cross-shelf [119]. The suspended sediment concentrations returned to pre-storm levels 2–3 days following the passage of the typhoon [119]. It is likely that much of the fine sediment resuspended by Hurricane Ian also settled to the benthos in the weeks between the storm and our post-hurricane sampling.

4.4. Microbial Community Changes

Surface water microbial communities changed after Hurricane Ian, but it is unclear to what extent the changes were due to the hurricane as opposed to regular seasonal variation, since the seasonality of microbial communities of the SWFS has not been characterized. Among the microbial community changes likely attributable to the hurricane are increases in prevalence of heterotrophs (Gammaproteobacteria and Bacteroidia) and the reduction of autotrophs (Cyanophyceae) in surface water, supported by picocyanobacteria counts. Similar changes were observed following a typhoon passing New Caledonia in the South Pacific [120]. This may reflect a shift in community metabolism towards heterotrophic decomposition, perhaps in response to the increased supply of complex and simple carbohydrates [121]. Given the high river discharges and run-off to the SWFS post-hurricane [67,68], large amounts of allochthonous dissolved organic carbon (DOC) were likely imported, prompting the shift in marine community composition. Extreme weather events can account for 20–50% of annual riverine DOC flux to coastal regions [79,122]. A shift from autotrophs to heterotrophs was also documented in the coastal waters of the Outer Banks, North Carolina after the hurricanes Florence and Michael sequentially hit the region in the 2018 hurricane season [123]. That study also found that at marine sites (>25 ppt) Flavobacteriia, Gammaproteobacteria, and unclassified Bacteroidetes all increased in relative abundance during the 2018 hurricane season when compared to communities from 2016 and 2019. They also saw a decrease in Alphaproteobacteria, as in this study. This could indicate an increase in nutrient concentrations as many marine Alphaproteobacteria are adapted to oligotrophic conditions [124,125].

Some of the taxa in sediments that increased post-hurricane (Actinobacteria, Clostridia, and Bacilli) are predominantly freshwater and terrestrially associated. Although Actinobacteria is widespread and can account for up to 13% relative abundance in other marine sediments, it can indicate human influence on coastal waters [126,127]. Some taxa among Actinobacteria, Clostridia, and Bacilli degrade woody recalcitrant lignin [128,129], and their increases in relative abundance may be in response to the large fluxes of lignin to the coast known to occur after major storm events [123]. The reduction of Planctomycetes and Bacteroidia may be due to the decline of fleshy macroalgae seen post-hurricane, as these microbes are important components of the biofilms of macroalgae [130,131]. Planctomycetes and sediment Bacteroidia are the typical anerobic mineralizers of marine sediments, responding to polysaccarides of algal origin that end up within the sediment [131,132]. Post-hurricane, these taxa may have been displaced by estuary-associated decomposers of more recalcitrant materials. Overall, the microbial community changes observed post-hurricane are consistent with an estuarization of the shelf by freshwater and terrigenous inputs.

5. Conclusions

Hurricane Ian significantly impacted the water quality, microbial communities, substrates, and benthic habitats of the SWFS through a combination of physical and chemical disturbances. The intense wave energy generated by the storm caused widespread scouring and sediment transport, leading to significant physical alterations in substrate availability and composition. Similarly, the massive influx of freshwater from coastal watersheds introduced large quantities of nutrients, DOC, and sediments, which triggered chemical changes that reshaped both water column dynamics and benthic ecosystem structure. These disturbances led to significant changes in microbial community composition and a dramatic decline in macrobenthic cover, with especially notable losses in habitat building groups like seagrass, stony corals, and calcareous algae.

While hurricanes are a natural part of the disturbance regime of the SWFS, anthropogenic pressures, including altered watershed hydrology, coastal development, and nutrient pollution, can synergistically worsen the effects [32,133,134]. In addition to better understanding the mechanisms of hurricane disturbance to shelf environments and identifying the potential to mitigate them through watershed management actions, there is a need to better understand the dynamics of recovery and resilience on the SWFS. Future research must prioritize understanding the resilience and recovery dynamics of benthic habitats across different substrate types, depth gradients, and proximity to riverine inputs. Patch dynamics, rates of ecological succession, and resistance/resilience of continental shelf habitats to multiple disturbances must be explored across the ecotones of the SWFS (e.g., hard and soft substrate habitats, nearshore, and offshore). Improved mapping of hard-bottom habitats and routine monitoring of water quality and benthic communities are essential for establishing baseline conditions, detecting early signs of degradation, and assessing the cumulative impacts of natural and anthropogenic stressors. E.g., one value of routine monitoring would be enabling assessment of how long post-disturbance ecosystem changes endure in this system. Given the expected increase in tropical cyclone intensity due to climate change, it is crucial to develop proactive plans to protect the biodiversity and health of the SWFS. These plans should include flexible management approaches, involve local communities, and support research across different fields to help maintain the resilience and long-term health of this important marine ecosystem.