Structure and Composition of Rhodolith Beds from the Sergipe-Alagoas Basin (NE Brazil, Southwestern Atlantic)

Abstract

Rhodolith beds are biogenic benthic habitats mainly formed by unattached, non-geniculate coralline algae, which can be inhabited by many associated species. The Brazilian continental shelf encompasses the largest continuous rhodolith bed in the world. This study was based on samples obtained from seven sites and videos taken by a Remotely Operated Vehicle (ROV) at four transects off the Sergipe-Alagoas Coast on the northeast Brazilian shelf. ROV operations and bottom trawl sampling revealed the occurrence of rhodolith beds between 25 and 54 m depths. At the shallower depths, fruticose (branching) rhodoliths (maërl) appear in troughs of ripples, and other non-branching rhodoliths occur associated with corals and sponge patches surrounded by bioclastic sand. Rhodoliths also occur in patches from 30 to 39 m depth; some are fused, forming larger, complex tridimensional structures. At deeper depths, from 40 to 54 m, the abundance of rhodoliths increases and occur associated with fleshy macroalgae on a smooth seafloor; some rhodoliths are fused into complex structures, locally some are fruticose (maërl), and others are partially buried by fine-grained sediment. The collected rhodoliths vary from fruticose in two sites to encrusting to lumpy, concentric and boxwork nodules in the rest; their size ranges from small (<1.5 cm) to large (~6 cm) and are mostly sub-spheroidal to spheroidal. A total of 16 red algal morpho-taxa were identified in the study sites. Two phases of growth can be distinguished in some rhodoliths by changes in color. The brownish inner cores yielded ages of 1600–1850 cal years before the present, whereas outer layers were much younger (180–50 years BP old). Growth layers appeared to have been separated by a long period of burial in the seafloor sediment. Other rhodoliths have ages of hundreds of years.

1. Introduction

Throughout their evolutionary history, non-geniculate coralline algae have been present in carbonate platforms as well as siliciclastic and mixed carbonate-siliciclastic systems [1]. Living and fossil non-geniculate coralline algae occur as crusts of varying thickness on both hard and soft substrates and free-living (unattached) nodules, also called rhodoliths and maërl, which can occur in large concentrations on the seafloor, known as rhodolith beds [2]. The rhodoliths occur worldwide from polar to tropical seas in a wide depth range on marine shelves [3,4] and are generally found where light is weak but sufficient for non-geniculate coralline algal growth and water motion and/or bioturbation are sufficient to move the rhodoliths [5]. Rhodoliths develop, in many cases, around a nucleus (bioclast or rock fragment) and can be composed of one or several non-geniculate coralline algal species, which can be intergrown with other encrusting organisms, such as bryozoans, gastropods, serpulids and encrusting foraminifera [6]. The rhodolith beds form three-dimensional complex biogenic benthic habitats [5] inhabited by a diverse biota of macroalgae and invertebrates [7,8,9], in particular crustaceans, polychaetes [10], cnidarians and chitons [11]. Some species of the associated fauna seem to be rhodolith-specific [12], influenced by rhodolith-forming species and rhodolith morphologies [13,14]. Furthermore, the phytobenthic community from the Sergipe-Alagoas (SEAL) Basin was characterized by a rich flora with a predominance of red algae (Rhodophyta) [15], showing an increasing gradient from the coast to the deepest regions, with lower values in the coastal areas under the influence of the São Francisco, Sergipe and Vaza-Barris rivers, and higher values in the deeper areas, characterized by the occurrence of more stable sediments [15], consisting of rhodolith beds.

Analyses of growth forms (e.g., lumpy, warty or encrusting) and the taxonomic composition of the non-geniculate coralline algae from the nucleus to the surface of the rhodoliths can provide valuable information on paleoenvironmental conditions during their development [1]. Rhodoliths and rhodolith beds do not grow continuously, as their growth could be interrupted by burial [16], sometimes making paleoenvironmental inferences from their growth record difficult [1].

The Brazilian continental shelf encompasses the largest continuous rhodolith beds in the world [17]. The fauna associated with rhodolith beds indicates high diversity of species across extensive areas of the continental shelf, but the main geographical gap of knowledge about rhodolith beds in Brazil is the northeast coast, specifically the continental shelf between Sergipe and Pernambuco, including the region of the São Francisco River mouth [18]. The marine sedimentation of the Sergipe-Alagoas continental shelf was analyzed in some studies that characterized Holocene sedimentation [19,20]. These studies show the spatial distribution of the main carbonate and siliciclastic components of the seabed [20,21] and their contributions to the formation of superficial sediment, comparing biotic and abiotic characteristics [22,23].

In this study, we aimed to (a) determine the distribution, structure and composition of rhodolith beds from the middle and outer shelf of the Sergipe-Alagoas Basin and (b) know their temporal evolution by establishing rhodolith ages obtained at different depths through carbon radiometric dating.

2. Study Area

The SEAL Basin on the northeast shelf of Brazil is subdivided into the Alagoas sub-basin, Sergipe sub-basin, Cabo sub-basin and Jacuípe sub-basin [24]. The study area (35° to 37° W and 10° to 11° S) is located within the Sergipe sub-basin and northern Jacuípe sub-basin, between the mouths of the Coruripe and Piaui-Real rivers (Figure 1). The Sergipe and south Alagoas continental shelf is considered shallow and narrow, with an average depth of 45 m and an average width of 28 km in Alagoas, 33 km in north Sergipe and 22 km in south Sergipe [23,25], with the shelf break beginning between 36 and 51 m depth in the center of the study area. It presents a gentle slope of approximately 1:570 and is characterized by a series of incised-shelf valleys, which have a northwest–southeast direction [23]. The main hydrographic basins of the study area are Coruripe, São Francisco, Japaratuba, Sergipe, Vaza Barris, Piauí and Real [26]. The continental shelf has different geomorphological settings along its 380 km length. These include five different areas, namely smooth areas or areas with little roughness, rough areas, canyons, incised valleys and aligned banks [27]. Rough areas are located south of Alagoas, between the São Francisco and Japaratuba rivers and south of Sergipe and correspond to the carbonate domain with calcareous algae [27]. We subdivided the carbonate domain into north, central and south sectors (Figure 1), which are separated by the two largest canyons, formed by the influence of the São Francisco and the Japaratuba rivers. The outer shelf of the north sector shows a reduced width and a soft and smooth relief in the southern part and a rugged and steeper relief in the northern one. The outer zone starts at the 30 m isobath and continues until the shelf break [23]. The incised valleys were partially filled with fine sediments, and carbonate deposition at the outer edge of the shelf is segmented by muddy zones adjacent to the river mouths. During the last deglaciation, the continental shelf of the SEAL was quickly flooded by the rapid rise in sea level. The shelf flooding started around 10,000 years BP and was completed 3000 years later [28,29]. The presence of sediments deposited over a geological history that extends from the penultimate marine regression represents the last marine transgression and records the stage of the current marine regression [21].

The tidal regime in the study area is semi-diurnal, with amplitude between 1.0 to 3.0 m [26], and the wave regime is high energy, predominating from NE and E in January to May (summer–autumn) and September to November (spring), and SE waves occur from March to August [26,30]. The normal sea surface temperature (SST) values vary between 25 and 28 °C and show a discrete decrease from north to south [26]. The SST in the summer and winter periods oscillates around 27–28 °C and 25–26 °C, respectively. Salinity (36–37‰) increases in the north and south, from the coastal region to the outer shelf, except near the river mouths, where salinity can reach values as low as 32–33‰ [26]. The primary productivity varies from 1.6–2.0 mg C/m²/day, and surface chlorophyll A concentrations vary from 0.7 and 5 mg/m³, with the highest values recorded in the coastal zone, which presented the potential of effective littoral drift with a predominant direction to the south [30]. The São Francisco River Basin discharges at the border between Sergipe and Alagoas and is the largest drainage basin in the Brazilian territory [29]. It is also one of the most regulated rivers in the country, and the discharge of the São Francisco River was drastically altered by the construction of six large dams built to generate energy and regulate flow [29]. Its average annual flow of more than 3000 m³/s declined by 80%. Its natural seasonal pulsation has been regulated to an almost constant flow of ~600 m³/s [25], and approximately 95% of the river’s total solids in suspension (TSS) are retained within the dam reservoirs [22,29,30]. The impact of the dams has considerably modified the main estuarine processes and sediment transportation; coastal erosion has also been drastically increased [31,32,33,34]. The São Francisco River and its delta have been featured in global compilations on the effects of sediment retention in dams and the risks represented by climate change [35]. Since 1960, rainfall has declined by approximately 27%, and average annual precipitation is projected to decrease up to 35% by 2080, compared to the average of 1961–1990, using the A1B scenario of the IPCC [35].

Currently, the continental shelf is characterized by low fluvial input and greater influence of the tropical surface waters (TSW) of the South Equatorial Current (SEC) that flows from east to west across the Atlantic Ocean [26,30]. In the proximity of the São Francisco and Japaratuba canyons during the summer months (December), there are lower temperatures and nutrient-rich waters due to the seasonal upwelling of the South Atlantic Central Water [26]. These environmental conditions promote the predominance of carbonate features from 30 m to 60 m depth, dominated by gravel and biogenic carbonate components [19,21,36] with an associated biota [19].

Coastal sediment transport along the Sergipe coast is 790,000 m³/year, with about 658,000 m³/year in the NE–SW direction and 132,000 m³/year in the opposite direction [37]. These processes influence the coastal region, either through coastal erosion in certain places or through deposition in others, as observed on the coast of the Sergipe State [38]. Large areas of smooth mud surfaces have been identified in front of the mouths of the São Francisco and Japaratuba rivers [39]. The middle and outer shelf, ca. 10,000 km² of the Sergipe–Alagoas shelf, are mainly characterized by carbonate deposits [39]. In southern Alagoas, these carbonate sediments are also found near the coastline [27]. Rough surfaces were found in the north sector, in the central sector, between the mentioned rivers, and in a larger area in the south sector. Siliciclastic components (quartz and rock fragments) are present along the inner shelf, up to 30 m depth, while carbonate contents that are predominate on the middle to outer shelf consist of rhodolith beds (Figure 1) that comprise non-geniculate coralline algae, foraminifera, bivalves and bryozoans (82%), while green algae Halimeda sp., echinoderms (spines and tests of urchins), spicules of sponges, fragments of crustaceans, gastropods, scaphopods, serpulids and corals are minor components, representing the remaining 18% [19,21].

3. Materials and Methods

3.1. Sampling

Data were collected within the framework of the SEAL Basin Regional Characterization Project carried out by Petrobras (Brazilian Petroleum S.A., Rio de Janeiro, RJ, Brazil) and the Federal University of Sergipe (UFS) from 2008 to 2018. A total of 53 high-definition video images, comprising 33 h of video inspections, was recorded from February to March 2014 [40] with a remotely operated vehicle (ROV) along four transects with a length of 4 km per transect, ranging from 25 to 46 m depth in the south sector of the study area (Figure 1, Supplementary Material Table S1 and Video S1). The cover percentage of rhodoliths was estimated by the proportion of the total video time spent on rhodolith concentrations.

Due to the high facies complexity of the continental margin of the Sergipe and Alagoas shelf and the distribution patterns of the biota along environmental gradients, the strategy for selecting the collection sites considered longitudinal, latitudinal and bathymetric gradients [41]. Rhodolith samples (404 rhodoliths) were obtained from two expeditions: the first carried out from 29 January to 6 February 2011 aboard the R/V Luke Thomas and the second from 24 June to 3 July 2011 aboard the R/V Seward Johnson, using a trawl [41]. The samples were stored in the Laboratory of Benthos from UFS and later ceded to the Instituto de Pesquisas Jardim Botânico do Rio de Janeiro (IPJBRJ). Rhodolith sampling sites and ROV video transects are shown in Figure 1. A total of seven sites between 27 and 54 m in depth were sampled along the middle and outer shelf of the SEAL Basin (Figure 1 and

Table 1. Sampling site locations, water depth and number (n = 404) of samples collected and analyzed in the present study. Coordinates are in SIRGAS 2000 Datum.

Site	Sector	Water Depth (m)	Latitude	Longitude	Samples (n)
1	North	30	−10.2789	−35.9792	07
2	North	30	−10.4875	−36.1245	173
3	Central	27	−10.8223	−36.6118	04
4	Central	47	−10.8694	−36.5472	141
5	South	47	−11.1642	−36.8839	05
6	South	54	−11.3389	−37.0556	18
7	South	50	−11.5989	−37.2384	56

). The geomorphological features of the shelf evaluated by high-resolution mapping in other studies [21,22], combined with the biofacies classes, sedimentological data [21] and the bathymetry dataset [22], were incorporated into a geographic information system (GIS) environment on the QGIS platform to produce the map in Figure 1.

3.2. Sample Characterization

The non-geniculate coralline algal growth forms were classified based on their morphology as encrusting, encrusting warty, lumpy and fruticose, according to [42,43]. The latter are monospecific rhodoliths lacking a macroscopic nucleus and characterized by a high prevalence of protuberances [44]. Based on their longest axes, rhodoliths were divided into large (d > 3 cm), medium (d = 1.5–3 cm) and small (d < 1.5 cm) size classes. Rhodolith sphericity was determined by measuring the longest, intermediate and shortest axes [45,46], and the triangular diagram plotting spreadsheet (TRI-PLOT) was used to separate ellipsoidal, discoidal or spheroidal shapes [46]. Species identification of crusts on the outermost surface of rhodoliths was conducted solely through the morpho-anatomical observation of histological sections (Supplementary Material Figure S1) of fertile material [44,47,48,49,50]. A permutational multivariate analysis of variance (PERMANOVA) with the software Primer (v.6) + PERMANOVA was used to evaluate possible significant differences in rhodolith size between the different sites and depths. Images of fruticose (branching) rhodoliths were used to derive a morphological parameter, the mean solidity index (degree of complexity of the thalli, which is reflected as values between 0—fully complex shape—and 1—no ramifications) of the thalli, corresponding to the surface of a given thallus of the fruticose rhodoliths on the image plane divided by its convex surface [51]. These analyses were performed using ImageJ 1.53 free software [51].

To characterize the inner structure of rhodoliths, 15 samples were selected and sectioned along their longest axis with a tungsten handsaw (Dremel 3000, Dremel, Bosch Power Tools). Both halves of the rhodoliths were then photographed using a digital camera (Nikon D7000, Nikon). The inner structure was defined as a boxwork, asymmetrical or concentric arrangement [42]. Small blocks (3 × 3 cm) were cut from 10 selected samples and were sent to the National Petrographic Service Inc. (Rosenberg, TX, USA) to make thin sections. The carbonate texture [52,53] was identified, and the Bioerosion Index (BI) was determined [54] to estimate the extent to which the primary features were bioeroded. Fossil marine macroinvertebrates, foraminifera and borings were identified to the most precise taxonomic level possible. The average proportions (%) of the main rhodolith components, trace fossils, voids and sediment were estimated by counting on images taken from each thin section (10 images per sample (n = 10), with 50 points randomly distributed on each image) using the CoralNet program ([55], https://coralnet.ucsd.edu/; accessed on 12 March 2021). Voids include constructional voids (intraskeletal and spaces among encruster skeletons) and bioerosion [56]. Voucher specimens of algae were deposited in the Herbarium of the Rio de Janeiro Botanical Garden (RB) (Supplementary Material Table S2).

To determine the radiocarbon age, eight subsamples of fossils in rhodoliths were selected. About 10–20 mg of carbonate powder was obtained from each subsample with a tungsten handsaw (Dremel 3000, Bosch Power Tools) and then chemically pretreated with HCl. They were analyzed at the Center for Applied Isotope Studies (University of Georgia, Athens, GA, USA) using an Accelerator Mass Spectrometer (AMS, nuPlasma II multi-collector). All conventional radiocarbon ages (between 603 and 50,779 years before present (BP)) were recalibrated using the software CALIB 8.2 ([57], http://calib.org/; accessed on 28 August 2020), using the Marine20 calibration curve [58] and a global reservoir effect. Dates are reported as calendar years before present (cal year BP; “present” = 1950 CE) with a 2-sigma confidence interval.

4. Results

4.1. Seabed Characterization in the Examined ROV Transect Lines

In the south sector (around site 5), living rhodoliths were always associated with fleshy macroalgae and were recorded along the four ROV transects (totaling 16 km in length, 4 km per transect). At the shallower depths (25–30 m) in this sector, fruticose (branching) rhodoliths (maërl) appeared in troughs of ripples on bioclastic sands (1.6 km; 10% of ROV transect length, Figure 2A,B). At the same depth range, patches of encrusting to lumpy rhodoliths (Figure 2C,D) were associated with corals, sponges and macroalgae (Figure 3A,C,D) and were surrounded by bioclastic sand occupying another 1.6 km (10%) of the transect length (Figure 2C,D). At depths between 30 and 39 m, rhodoliths patches extended for 4.8 km (30%) of the transect length and included small fused rhodoliths creating larger complex 3D structures together with corals and sponges (Figure 2E) associated with fleshy macroalgae (Figure 3B). From 40 to 54 m, the abundance of rhodoliths increased. They occurred in association with fleshy macroalgae on a smooth seafloor (6.4 km; 40% of transect length, Figure 2F), and some include fused rhodolith structures (Figure 2G). At this depth interval, 5% (0.8 km) of the transect length was occupied by rhodoliths partially buried by fine-grained sediment (Figure 2H). Fruticose (branching) rhodoliths (maërl) also occurred locally at 47 m. A reef fish community was observed associated with rhodolith patches between 36 and 44 m depth, including Holocentrus adscensionis, Gymnothorax moringa and Cephalopholis fulva (Figure 2E, and Figure 3B,E,F) as the main species. Massive sponge species were recorded attached to rhodoliths, including Aplysina sp. and Ircinia strobilina (Figure 2D and Figure 3C). The videos showed differences in rhodolith distribution between the northeast (green colour in Figure 1) and southwest (red colour in Figure 1). Rhodoliths were sparsely grouped, and some of them fused (Figure 2A–D) in the former transect, whereas rhodolith patches were larger with larger aggregates of fused rhodoliths (Figure 2E–H) in the latter.

4.2. North Sector

The rhodoliths collected in the middle shelf at 30 m depth from site 1 were large (5.7–20.4 cm at the longest axis and mean diameter that ranged from 4.2–14.3 cm; Figure 4A and Figure 5A), encrusting and encrusting warty with mostly sub-discoidal (42.8%) to sub-spheroidal (28.5%) shapes (Figure 5B and

Table 2. Rhodolith size, coralline algal growth-form, rhodolith shape (%) and rhodolith-forming coralline algal species from the SEAL Basin. Surface (x-blue); outermost portion (x-green); innermost portion (x-red).

Site	Sector	(n)	Size			CCA Growth-Forms				Rhodolith Shapes (%)						Rhodolith-Forming Coralline Algae Species
			L	M	S	EN	EW	FR	L	SP	SSP	EL	SEL	DI	SDI	H	HR	HC	Hsp	Lisp	PF	Spsp	LtP	Ltsp	Lsp	Rs	ME	Msp	SFr	Ssp	Psp
1	N	7	7	0	0	4	1	0	2	14.29	28.57	0	0	14.29	42.86	x		x	x					x				x		x
2	N	173	3	1	169	2	0	169	2	75	25	0	0	0	0		x							x	x	x	x	x
3	C	4	4	0	0	4	0	0	0	0	50	0	0	0	50							x									x
4	C	141	0	0	141	2	7	130	2	0	0	0	0	0	0									x	x					x
5	S	5	4	1	0	4	0	0	1	60	40	0	0	0	0	x	x		xx		x		x		x				x	xx	xx
6	S	18	18	0	0	11	1	0	6	22.22	38.89	5.56	16.67	0	16.67				x	x		x	x	x	xx	xx				x	x
7	S	56	54	2	0	36	8	0	12	39.29	53.57	0	1.79	0	5.36	x			x	xx		x	xx	xx	xx	x	x			x	xx

N = North sector; C = Central sector; S = South sector; L = Large (d > 3 cm); M = Medium (d = 1.5–3 cm); S = Small (d< 1.5 cm); EN = Encrusting; EW = Encrusting warty; FR = Fruticose; L = Lumpy; SP = Spheroidal; SSP = Sub-spheroidal; EL = Ellipsoidal; SEL = Sub-ellipsoidal; DI = Discoidal; SDI = Sub-discoidal; H = unidentified Hapalidiales; HR = Harveylithon gr. rupestre; HC = Harveylithon gr. catarinense; Hsp = Harveylithon sp.; Lisp = Lithoporella sp.; PF = Pneophyllum gr. fragile; Spsp = Spongites sp.; LtP = Lithophyllum gr. pustulatum; Ltsp = Lithophyllum sp.; Lsp = Lithothamnion sp.; Rs = Roseolithon sp.; ME = Melyvonnea erubescens; Msp = Mesophyllum sp.; SFr = Sporolithon franciscanum; Ssp = Sporolithon sp.; Psp = Peyssonnelia sp.

). The rhodoliths examined from this site had living encrusting coralline algae (Harveylithon sp., Sporolithon sp.) and sponges partially covering their surfaces (Figure 4A). The nuclei consisted of coral fragments (Figure 6A) or other bioclasts. Coralline algae intergrown with bryozoans, corals and serpulids formed an asymmetrical cover with a boxwork structure. Coralline algae included Mesophyllum sp., Lithophyllum sp., Harveylithon gr. catarinense (I.O.Costa, P.A.Horta & J.M.C.Nunes 2019) and an unidentified Hapalidiales species (Figure 6C and Table 2). The three rhodoliths examined were heavily bored (Bioerosion Index is 4) by sponges (Entobia) and bivalves (Gastrochaenolites). Constructional voids and bioerosion traces varied from empty to partially filled by wackestone rich in sponge spicules and micrite-rich packstone. Geniculate coralline algae, benthic and planktonic foraminifera, and solitary corals appeared in the fillings (Figure 6B and

Table 3. Main components (%) in ten analyzed rhodolith samples from the SEAL Basin. 1gp = first growth phase (innermost portion); 2gp = second growth phase (outermost portion).

	North Sector (Site 1)			Central Sector (Site 3)	South Sector (Site 5)		South Sector (Site 6)			South Sector (Site 7)
Sample number	1	2	3	4	5	6	7	8-1gp	8-2gp	9	10-1gp	10-2gp
Bryozoan	3.94	12.29	7.54	0.58	8.4	5.62	47.4	17.26	37.27	16.38	26.25	7.75
Coral	20.15	0	0.91	52.15	2.31	8.51	0	0	0	0	0	0
Geniculate coralline algae	1.34	0	0	0	0.8	0.35	0	0	0	0	0	0
Non-geniculate coralline Algae	19.28	56.53	48.36	6.92	31.36	52.7	21.46	23.47	32.38	38.19	28.75	59.41
Echinodermata	2.13	0.45	0.88	0	0	0	0	0	0	0	0	0
Foraminifera	0.42	0.44	0.46	0	1.38	0	0	0.39	0.71	0	0	0.64
Gastropod	0	0	0.24	0.1	0	0	0	13.4	0.36	9.93	7.5	1.57
Serpulid	0	1.67	0	0.83	0.83	0.6	0	0	0.7	0.42	0	0
Sponge spicules	0	0	0.54	3.44	0	5.03	0	0	0	0	0	1.77
Filling sediments	33.16	6.69	5.54	9.15	13.02	5.36	1.9	6.28	3.68	5.15	17.5	9.95
Siliciclastic grains	5.48	2.23	0	3.68	4.65	0	1.83	2.14	0	0	0	0
Borings	13.7	24.17	16.25	18.06	20.6	32.5	11.67	19.6	17.5	16.67	17.5	17.3
Constructional voids	0	0	9.51	0.91	0.9	0.42	14.47	16.32	4.71	11.04	2.5	0.21

). Radiocarbon dating of a coral in the innermost part of a rhodolith sample yielded an age of 789 cal yr BP (Figure 7A), and a non-geniculate coralline algae in the outermost part of another rhodolith sample yielded an age of 180 cal yr BP (Figure 7B and

Table 4. Radiocarbon age of different components in rhodoliths collected at different sites from 3 sectors (N, North; C, Central; and S, South) of the SEAL Basin. Coordinates are in SIRGAS 2000 Datum. * Values < 433 yr BP cannot be calibrated at the curve Marine20; BP = before present.

Site	Sector	Latitude	Longitude	Water Depth (m)	Dated Component	Conventional Age (BP)	Calibrated 14C Age (Cal yr BP)	95% Range (cal BP)
1	N	−10.2789	−35.9792	30	Coral	1840 ± 200	789	666–922
1	N	−10.2789	−35.9792	30	Coralline algae	180 * ± 200	-	-
3	C	−10.8223	−36.6118	27	Coral	380 * ± 250	-	-
5	S	−11.1642	−36.8839	47	Coralline algae	340 * ± 200	-	-
6	S	−11.3389	−37.0556	54	Coralline algae	760 ± 200	1825	1700–1950
7	S	−11.5989	−37.2384	50	Coralline algae	50 * ± 200	-	-
7	S	−11.5989	−37.2384	50	Gastropod	4040 ± 250	1851	1690–2014
7	S	−11.5989	−37.2384	50	Bryozoans	960 ± 250	1610	1476–1755

).

Rhodoliths from site 2 were mostly fruticose (branching) less than 2 cm in length (Figure 8A). They exhibited complex shapes with branches varying from small to medium (Solidity Index = 0.386–0.937). Together with bryozoans, specimens of Harveylithon gr. rupestre, Roseolithon sp., Lithothamnion sp., Lithophyllum sp., Melyvonnea gr. erubescens and Mesophyllum sp. were identified at the rhodolith surfaces (Figure 8A and Table 2). Larger rhodoliths with asymmetrical boxwork to the concentric inner arrangement also occurred. They were medium to large (2.1–13.7 cm; Figure 4B and Figure 5A), encrusting and lumpy with mostly spheroidal (75%) shapes (Figure 5B and Table 2). Rhodolith were formed mainly by non-geniculate coralline algae with bryozoans and corals as secondary components.

4.3. Central Sector

Rhodoliths from site 3 at 27 m depth were large in size (4–9.5 cm at the longest axis and mean diameter that ranged from 3.0-6.8 cm; Figure 4C and Figure 5A), encrusting with sub-spheroidal (50%) and sub-discoidal (50%) shapes (Figure 5B and Table 2). Four rhodoliths examined had surfaces partially covered by red algae (Spongites sp., Peyssonnelia sp.). They were characterized by a highly asymmetrical boxwork inner arrangement, with nuclei formed mostly by corals. The covers consist of non-geniculate coralline algae, serpulids, bryozoans and encrusting foraminifers (Figure 9A,B and Table 3). Both nuclei and covers are bored by sponges (Entobia), and the Bioerosion Index is 3. The borings were partially filled by wackestone and packstone rich in micrite, with bioclasts of echinoids and mollusks, sponge spicules, and siliciclastic grains, mainly quartz (Figure 9C). Radiocarbon dating of a coral nucleus of a rhodolith yielded an age of 380 yr BP (Figure 7C and Table 4).

The rhodoliths collected in the outer shelf at 47 m depth from site 4 were mainly fruticose, and a few were small encrusting warty. Fruticose rhodoliths were less than 3 cm in size (Figure 8B), with low complexity, i.e., exhibit medium branches and no branching (Solidity Index = 0.614–1.000). They were mainly made up of Lithothamnion sp. and Lithophyllum sp., whereas bryozoans were secondary components. Sporolithon sp. was identified at the surface of small encrusting warty rhodoliths (Figure 8B).

4.4. South Sector

Rhodoliths from site 5 at 47 m depth were medium to large (2.7–10.3 cm at the longest axis and mean diameter that ranged from 2.4-8.8 cm; Figure 4D and Figure 5A), encrusting, encrusting warty, and lumpy with mostly spheroidal (60%) shapes (Figure 5B and Table 2). The rhodolith surfaces were partially covered by living encrusting red algae (Harveylithon gr. rupestre, Harveylithon sp., Pneophyllum gr. fragile, Sporolithon franciscanum, Sporolithon sp. and Peyssonneliaceans, Figure 4D) and Cliona cf. schmidtii sponges. The rhodoliths had an asymmetrical boxwork inner arrangement, formed mostly by non-geniculate coralline algae, bryozoans, encrusting foraminifera and serpulids. Lithothamnion sp. and unidentifiable Hapalidiales, Harveylithon sp., Lithophyllum gr. pustulatum, thin unidentifiable laminar thalli, Sporolithon sp. and Peyssonneliaceans are the algal components (Table 2). The skeletal framework was heavily bored by sponges (Entobia), and the Bioerosion Index is 4. Intraskeletal and constructional voids, as well as sponge borings (Figure 10A and Table 3), varied from empty to partially filled by wackestone to packstone. Geniculate coralline algae, benthic and planktonic foraminifera appear with echinoid and mollusk fragments and siliciclastic grains, mainly quartz, in the fillings. Sponge spicules are common in the empty spaces in the outer part. Radiocarbon dating of non-geniculate coralline algae in the middle part of a rhodolith yielded an age of 340 yr BP (Figure 7D and Table 4).

Rhodoliths from site 6 at 54 m depth were large (3.3–13.8 cm at the longest axis and mean diameter that ranged from 2.6–7.6 cm; Figure 4E and Figure 5A), encrusting, encrusting warty, and lumpy with sub-spheroidal (38%) and spheroidal (22%) shapes (Figure 5B and Table 2). The surfaces were partially covered by Peyssonneliaceans, Roseolithon sp. and thin unidentifiable non-geniculate coralline algae. The examined rhodoliths consist of brownish bioclastic nuclei, commonly bryozoan nodules or gastropod shells (Figure 4D, Figure 7E,G and Figure 10B,C), and covers made of non-geniculate coralline algae intergrown with bryozoans with an asymmetrical concentric arrangement. Serpulids and encrusting foraminifers were secondary framework builders. Roseolithon sp., Lithothamnion sp., Spongites sp., Harveylithon sp., Lithoporella sp., Lithophyllum gr. pustulatum, Lithophyllum sp., Sporolithon sp., unidentifiable thin laminar thalli and Peyssonneliaceans are the algal components (Table 2). The rhodoliths were slightly bioeroded (Bioerosion Index is 2) by sponges and worms. Most intraskeletal and constructional voids and borings are empty (Figure 10B), with a small proportion partially filled by wackestone and packstone rich in micrite with siliciclastic grains, mainly quartz (Table 3). Radiocarbon dating of non-geniculate coralline algae in the middle part of a rhodolith yielded an age of 1825 cal yr BP (Figure 7E and Table 4).

Rhodoliths from site 7 at 50 m depth were medium to large (3.4–14.5 cm at the longest axis and mean diameter that ranged from 2.3–11.7 cm; Figure 4F and Figure 5A), encrusting, encrusting warty, and lumpy with mostly sub-spheroidal (53%) shapes (Figure 5B and Table 2). The surfaces are partially covered by thin crusts of Melyvonnea gr. erubescens, Lithothamnion sp., Lithophyllum gr. pustulatum and Peyssonneliaceans (Table 2). Most rhodoliths show two distinct phases of growth. The inner phases, brownish in color, are made of bioclastic nuclei, such as gastropod shells and bryozoan fragments, or frameworks of bryozoans, serpulids and small oysters with a small proportion of non-geniculate coralline algae, including Lithophyllum sp. and Lithoporella sp. (Figure 10C and Table 3). These inner phases are bored by sponges. Borings and intraskeletal voids are partially filled by mudstone to packstone. The outer phases consist of non-geniculate coralline algae intergrown with bryozoans and minor serpulids and encrusting foraminifers with an asymmetrical concentric arrangement. Roseolithon sp., Lithothamnion sp., unidentifiable Hapalidiales, Spongites sp., Harveylithon sp., Lithoporella sp., Lithophyllum gr. pustulatum, Lithophyllum sp., Sporolithon sp., unidentifiable thin laminar thalli and Peyssonneliaceans are the algal components (Table 2). The outer and inner phases, including the boring fillings, are bioeroded by sponges, and most of the voids are empty. A gastropod shell and a bryozoan in the inner phases of two rhodoliths were dated as 1851 and 1610 cal yr BP, respectively. Radiocarbon dating of non-geniculate coralline algae from the outer phase of another rhodolith yielded an age of 50 years BP (Figure 7F–H and Table 4).

4.5. Comparison of Rhodolith Distribution and Characteristics among Sectors

According to the biofacies distribution at SEAL (See Section 2 Figure 1), the different bottom features of the rhodolith beds in the south sector of the SEAL Basin, observed in ROV video, extend to the middle and outer shelf of the north and central sectors.

The size of rhodoliths differs among sites (PERMANOVA; F = 6.8764; p = 0.0002). A posteriori test, pairwise, showed that rhodoliths of site 1 were larger than those of deeper sites 6 (p = 0.0014) and 7 (p = 0.0002) (Figure 5A). Small fruticose rhodoliths (maërl) occurred both in ripple troughs in shallow settings in the south sector and the deeper sampling sites in the north and central sectors. The rest of the rhodoliths exhibited concentric and boxwork inner arrangements made up of encrusting to lumpy coralline algae in all sites. In the south sector, however, the proportion of constructional void is higher, and a concentric arrangement was prevalent in rhodoliths at the two deepest sampling sites. In these sites, a two-phase development of analyzed rhodoliths was observed. The nuclei of shallowest rhodoliths are commonly formed by corals, which are absent in deeper rhodoliths from the south sector.

The number of coralline morpho-taxa involved in rhodolith formation was higher in the south sector than in the other two. The nuclei of rhodoliths in the north and central sectors include coral fragments, which are absent in the south sector, where the nuclei are formed either by mollusk shells or bryozoan skeletons.

5. Discussion

The rhodoliths analyzed in this study started to grow between 1.9 and 1.8 ka cal BP, a long time after the Holocene flooding of the SEAL shelf [35]. The bathymetry of the investigated rhodolith beds lies within the large range of rhodolith beds in Brazil, which occur between 12 and 95 m depth [59,60,61,62,63,64,65]. Temperature, nutrients, sedimentation and water current velocity are considered the main environmental drivers of rhodolith development in the Brazilian ecoregions [65]. Specifically, sediment influx depends on discharge from large rivers and directly influences the distribution of bottom sediments and rhodolith beds [59]. On the inner SEAL shelf, rhodolith growth is inhibited by riverine sediment influx (mud/sand siliciclastic sediments), but rhodolith beds can occur locally (See Section 2 Figure 1). Rhodolith beds on the SEAL shelf were concentrated north and south of the São Francisco river mouth. A similar distribution can be observed in the Rio Doce shelf, where rhodoliths are most abundant in two discontinuous areas at depths between 45 and 65 m. The area south of the Rio Doce mouth is larger and shows higher rhodolith cover, while the area north of the mouth shows very low rhodolith cover (<10%), which was interpreted as a consequence of long-term deposition of fine sediments [59].

Rhodolith beds on the SEAL shelf are widespread in the middle-outer shelf, following the same trend observed along the Brazilian margin, but their distribution is interrupted by the presence of shelf-incised valleys and canyons, such as the Sao Francisco canyon, and minor incisions dissecting the shelf (Figure 1). In narrow mixed continental shelves, such as the SEAL shelf, transitions from terrigenous sediments to carbonates occur laterally due to lateral variation in environmental controls, such as sediment input and carbonate production [66]. Moreover, the presence of canyon systems in narrow shelves favor the possibility of discontinuities in sediment distribution even during high stands [67]. Shelf-incised valleys, canyons and riverine plumes led to the patchy distribution in large-scale rhodolith beds on the SEAL shelf (Figure 1). On wider shelves, such as the Abrolhos shelf, which also has low sediment input, rhodolith beds are extensive and more continuous [63,66]. At a smaller scale, within individual rhodolith beds mapped on the SEAL shelf (Figure 1), the distribution of rhodoliths is patchy with varying rhodolith density. This uneven distribution has been reported for deep rhodolith beds and seems to be mainly related to seafloor morphology and/or bottom currents [59,68,69,70].

The video surveys confirmed the occurrence of patches of fused rhodoliths in depths between 30 and 50 m, creating complex large three-dimensional structures with up to 10% of cover (Figure 2F and Figure 3F). Similar structures occur in deeper waters in other regions in Brazil, i.e., from around 50 to 80 m depths on the Rocas Atoll shelf [69] and from 65 to 105 m on the Doce River shelf [59]. On the latter shelf, carbonate concretions, including rhodoliths, occur in areas with a high density of rhodoliths, suggesting that the concretions were formed by rhodolith fusions, later overgrown by coralline algae and other encrusters, mainly bryozoans [59]. In the Mediterranean Sea, the so-called coralligenous de plateau (tridimensional build-ups of calcareous encrusting algae growing on flat surfaces) are common on almost all shelves, at low irradiance levels and in relatively calm waters [71,72]. In some cases, they were interpreted as frameworks developed from a coalescence of rhodoliths [71,73,74], and, therefore, the patches of fused rhodoliths on the SEAL shelf can be considered initial steps of the formation of frameworks similar to coralligenous banks. However, fossil examples from the early Miocene in Sardinia [75] and the late Miocene in SE Spain [76] showed that fused rhodoliths do not necessarily promote further development of coralline algal frameworks.

The shallowest rhodoliths (27 m depth) in the north and central sectors commonly have large nuclei of coral fragments, which is in accordance with the occurrence of corals in the rhodolith patches observed by the ROV between 25 and 30 m depth. Similar nuclei characterize the rhodoliths collected at 20 m depth in the Queimada Grande Island [64] and at 23 m depth in the southern zone of the Amazon continental margin, which has the lowest riverine influx in the region off the Amazon mouth [62]. In the north sector of the SEAL shelf, the coral nuclei have an age of nearly 800 years BP, while the dated coralline algae cover is only a few hundred years old.

Beds of small fruticose rhodoliths (maërl) are uncommon on the Brazilian shelf. The fruticose rhodoliths in sites 2 and 4 exhibit low to medium branching or no ramification in some cases, which corresponds to low habitat complexity. The development and distribution of maërl are controlled by various environmental factors that remain difficult to discern [77,78]. Maërl beds can be shaped by the swell to form megaripples, but wave energy can rapidly limit the vitality of the maërl-forming species in contrast to encrusting coralline algae that can live in highly hydrodynamic environments [78]. The mobility of branching rhodoliths depends on wave agitation and currents, and they may only move occasionally due to bioturbation and severe storms [79]. If hydrodynamic conditions are too low, fine sediment particles can stifle algal growth [80], whereas under strong energy, the fine sediments are re-suspended, inhibiting photosynthetic activity, and coarser particles such as sand or broken shells bury the thalli [78]. Well-developed maërl beds occur in areas swept by currents, preventing the burial of rhodoliths by fine-grained sediment [42,81], although they also occur in areas with no significant energy but with reduced sediment influx [82]. Maërl beds are consequently very sensitive to sediment input and sediment mobility. On the north Atlantic coasts, maërl beds usually occur in shallow and sheltered bays [77,78]. In contrast, in the Spencer Gulf in South Australia, fruticose rhodolith pavements develop at depths of several tens of meters under the influence of moderate to high-energy currents of tidal origin that remove fine sediments [83]. In the Mediterranean Sea, fruticose rhodoliths are found in deeper settings, from 30 to 100 m [81,84,85,86]. In fossil examples, fruticose rhodoliths also tend to occur in paleoenvironments deeper than those of nodules formed by encrusting to lumpy thalli [87]. The occurrence of the two beds of maërl in the deeper sites of the north and central sectors of the SEAL seems to be favored by calm conditions in deep shelf settings sheltered from sediment influx. In the south sector, however, fruticose rhodoliths occur in higher energy areas in the troughs of ripples.

The concentric arrangement and high proportions of constructional void in rhodolith covers in sites of the south sector of the SEAL shelf are common in shallow-water rhodoliths in the south Espirito Santo and Rio Doce shelves [59,63]. High proportions of constructional voids were interpreted as reflecting low energy conditions [63], whereas concentric arrangements with little destruction by bioerosion indicate relatively short residence times of rhodoliths on the seafloor before burial. In the deepest sites of the south sector, a two-phase development of rhodoliths was observed. Nuclei are highly bored by sponges, whilst bioerosion in the covers is very low. In these rhodoliths, there is a large difference in age between growth phases: the inner phases yielded ages of 1600–1850 cal years BP, whereas the dated outer phase was only 50 years BP. The difference in bioerosion density and age can be explained by two phases of growth separated in time by a long period of burial in the seafloor sediment. Such a burial of rhodoliths by fine-grained sediments was recently observed in the ROV transects at site 5, between 40 and 46 m depths. A similar two-phase development of rhodoliths was observed in the central zone of the Amazon River mouth, at 50–55 m depth [62]. In this location, the growing gap was much shorter, about 400 years from the oldest age of the darkened interior [62]. The burial by sediment, subsequent exhumation and regrowth are probably related to variations in riverine discharge to the shelf [62]. Some deep-water rhodoliths of the Abrolhos Continental Shelf also show multiphasic growth with growth gaps of thousands of years caused by temporary burial and exhumation [63]. Rhodoliths with well-differentiated growth phases have also been described in different regions [88,89,90,91,92], with time gaps between growth phases of several hundred years. Multistory rhodoliths with two to several growth phases have also been described in the fossil records [1,93], suggesting that temporary burial is a common phenomenon in rhodolith beds.

The 15 rhodolith-forming coralline taxa recorded in this study are commonly found in the beds of northeast and southeast Brazil [18,59]. Among the identified taxa, the multispecific genus Roseolithon (previously recognized as the single species Lithothamnion crispatum in Brazil) occurs on surfaces and inner portions of rhodoliths from sites in the north and south sectors of the SEAL Basin. This is considered one of the main rhodolith-forming non-geniculate coralline algae of the Brazilian continental shelf [17,18,94]. The number of morphoanatomically identified taxa in the SEAL shelf is similar to that of the Abrolhos continental shelf (14 taxa) but substantially higher than numbers recorded in the Rio Doce shelf (10 taxa), Campos Basin (2 taxa), Amazon continental margin (6 taxa) and Fernando de Noronha Archipelago (6 taxa) [17,18,48,62,95]. However, all these taxon accounts based on morpho-anatomical identifications do not reflect the higher richness of phylogenetic species of coralline algae delimited in the different ecoregions of the Brazilian shelf by DNA-based methods [65]. In particular, in the northeastern Brazil ecoregion to which the SEAL shelf belongs, 23 phylogenetic species of coralline algae have been identified [65].

The extension and composition of the SEAL rhodolith beds show the importance of these ecosystems on the Brazilian continental shelf. This region has great economic importance due to its extensive areas of carbonate accumulation [22] and is subjected to multiple disturbances as direct and indirect consequences of human activities, such as mineral extraction of calcium carbonate for commercial use, bottom trawl fishing, and oil and gas extraction [21,96]. Knowledge about rhodolith beds is becoming even more urgent to support local conservation and management actions and assess the environmental impacts that the exploitation of its resources can generate on this rich marine ecosystem.

6. Conclusions

Rhodolith beds on the SEAL shelf mainly occur on the middle-outer shelf, at depths between 25 and 55 m. Sediment influx from the major rivers in the area and incision of the shelf by submarine canyons determine the large-scale areal distribution of rhodolith beds on the shelf. The spatial variability of rhodolith beds reveals living rhodoliths grouped in patches interspersed with bioclastic sediments and associated with fleshy macroalgae and a reef fish community; some rhodoliths are fused, forming coralline algal concretions; fruticose rhodoliths (maërl) occur both in ripple troughs in shallow waters and deeper sites of the north and central sectors. Coralline algal identification recorded a total of 15 taxa in the study sites, where some sites supported up to 11 different morphospecies; some species occur at five sites along the three sectors of the SEAL shelf. The rhodoliths from the north and central sectors commonly have large nuclei of coral fragments hundreds of years old. In the south sector, the rhodoliths have old nuclei (about 1800–1600 cal years BP) of gastropod and bryozoan fragments or coralline algae-bryozoan-serpulid boundstones. In this sector, rhodoliths show two growth phases with a difference in age of more than 1500 years, which can be explained by a long period of burial in the seabed sediment and later exhumation and resuming of growth. Increased knowledge about the role of these rhodolith beds in the benthic ecosystems is important to provide support for marine spatial planning programs, which could lead to more sustainable use of these habitats.