Building a baseline: a survey of the composition and distribution of the ichthyofauna of Guanabara Bay, a deeply impacted estuary

Abstract Biodiversity baselines are essential subsidies to evaluate how environmental changes and human impacts affect the special and temporal patterns of communities. This information is paramount to promote proper conservation and management for historically impacted environments such as Guanabara Bay, in southeastern Brazil. Here, we propose an ichthyofaunal baseline for this bay using gathered past data from 1889 to 2020, including literature records, scientific collections, biological sampling, and fisheries landing monitoring. A total of 220 species (203 teleosts and 17 elasmobranchs), distributed in 149 genera (136 teleosts and 13 elasmobranchs) and 72 families (61 teleosts and 11 elasmobranchs) were recorded, including the first record of a tiger-shark, Galeocerdo cuvier, in Guanabara Bay. Although the employed sampling effort was sufficient to represent the ichthyofauna in the middle and upper estuary, the Chao2 estimator indicates an even greater richness regarding the bay as a whole. Evidence of reduced abundance and probable local extinction over the decades was found, supporting the importance of implementing management and conservation strategies in the area. The ichthyofaunal distribution analyses revealed that areas close to conservation units are richer compared to their surroundings, indicating that this is an effective strategy to mitigate human impacts in the bay.


INTRODUCTION
Estuaries are highly dynamic coastal environments that exhibit a wide range of salinity, nutrient, and temperature variations, providing habitats, resources, and shelter to a variety of species at different life cycle stages (Silva-Junior et al., 2016;Wolanski, Elliott, 2016).Estuaries function as important nursery and feeding areas (Corrêa, Vianna, 2015;Santos et al., 2015;Andrade et al., 2016;Mérigot et al., 2017;Gonçalves-Silva, Vianna, 2018b), which are essential for the maintenance of several marine fish stocks (Santos et al., 2020).Even though these environments are known to contain few strictly resident species (Andrade-Tubino et al., 2008;Vianna et al., 2012;Silva-Junior et al., 2016;Gonçalves-Silva, Vianna, 2018a), their ichthyofaunal diversity displays a rich taxonomic composition, including many species of economic interest and others at serious risk of extinction.
The Guanabara Bay is the second largest Brazilian estuary, located in the metropolitan region of the state of Rio de Janeiro, presenting significant historical, environmental, touristic, and scenic importance.The bay also comprises an essential part of Rio de Janeiro's economy, since it harbors a major port area and supports the most productive estuarine fisheries in the region (Prestrelo, Vianna, 2016).Guanabara Bay has historically suffered from a series of human impacts associated to huge solid waste, untreated domestic sewage, and persistent pollutant inputs, such as metals and hydrocarbons (Pereira et al.,

MATERIAL AND METHODS
Study area.The Guanabara Bay (22°59'02.20''S-22°40'23.66''S;43°01'26.53''W-43°17'26.08"W) is a semi-enclosed tropical estuary located on the southeastern coast of Brazil, in the state of Rio de Janeiro, covering 384 km 2 , with an average volume of 1.87 x 10 9 m 3 of water, and a 4,080 km 2 drainage basin with maximum depth of 50 m in the central channel (Meniconi et al., 2012;Silva-Junior et al., 2016).It is characterized by seasonal salinity variations influenced by a connection with oceanic waters, the local rainfall regime, and tides.During the low rainfall period (June to August), the water column is more homogeneous, with little temperature and salinity variations, becoming vertically stratified during the rainy season (December to March), with the appearance of upwelling areas due to the penetration of the South Atlantic Central Water (SACW) that enters the estuary through its saline wedge (Valentin et al., 1999;Silva-Junior et al., 2016).The bay is categorized into three compartments (sensu Silva-Junior et al., 2016;Souza, Vianna, 2022): (i) the lower estuary, corresponding to the central channel and its banks, comprising the area suffering the greatest influence of the oceanic waters that enter the bay; (ii) the middle estuary, consisting of an intermediate transition area between the more saline waters of the lower estuary and the more brackish waters of the upper estuary, and (iii) the upper estuary, the innermost bay region under greater influence of continental waters from the local hydrographic basin.
The Guanabara Bay entrance was defined as the shortest distance between the east and west coasts (limit line, from the point of Forte São José, 22°56'24.41"S43° 09'06.66"W to the point of Fortaleza de Santa Cruz da Barra, 22°56'16.97"S43°08'06.30"W).Therefore, all records external to this line were considered as outside the estuarine region and were not included in our inventory.The bay was also divided into quadrants using the Quantum GIS (QGIS) software version 3.16.5 according to the same grid applied by the fishing landing monitoring efforts in Guanabara Bay (Prestelo, Vianna, 2016) (Fig. 1).Data compilation.Different strategies were employed to gather ichthyofaunal records in the Guanabara Bay.First, we made a compilation of scientific literature concerning the bay's ichthyofauna.A scientometric analysis was carried out at the Web of Science, SciELO and Scopus portals, covering articles from all available years, i.e., from 1921 to March 23, 2021.The search method applied two keyword fields linked by the connectors "AND" and "OR", the first referring to the study location (Guanabara Bay) and the second to the study group (ichthyofauna) (Tab.1).We added to the scientometric analysis results other published articles that were previously known by the authors.Then, data from two sets of fish samplings carried out by BioTecPesca/ UFRJ were added to the database.These bottom trawl samplings were carried out from 2005 to 2007 at quadrants C3, C5, D5, D7, E3, E5 and E7; and from 2013 to 2015 at quadrants C5 and D5.In addition, the records of species identified in two artisanal fishing landing monitoring programs at Guanabara Bay based on different commercial fishing gear (also carried out by BioTecPesca/UFRJ) were considered, the first in 2009 and 2010, and the second in 2013 and 2014.
Historical records were obtained from the online databases of the fish collections of Museu de Zoologia da Universidade de São Paulo, São Paulo (MZUSP) (records available online at https://mz.usp.br/pt/laboratorios/ictiologia/accessed on July 30, 2020) and the Museu Nacional, Universidade Federal do Rio de Janeiro, Rio de Janeiro (MNRJ) (records available online at https://ipt.sibbr.gov.br/mnrj/resource?r= mnrj_ ictiologia, accessed on July 30, 2020).In addition, we included to our data compilation the listings made by the BioTecPesca/UFRJ research group deposited at the Coleção de Peixes do Instituto de Biodiversidade e Sustentabilidade, Universidade Federal do Rio de Janeiro (NPM).The SpeciesLink Network (http://www.specieslink.net/)was used as another tool to compile records from several scientific collections."Guanabara" was employed as keyword and the records were filtered by taxon to include only fishes, considering all reports until July 21, 2020.All lot numbers of species included in our database are available in Tab.S1.
Regardless of the strategy employed, records were considered only at the species level, with species without previous confirmed records in the state of Rio de Janeiro considered as doubtful records and not included in the baseline.Records at genus or family level were also not considered.Taxonomic classification (at species level and above) and species known distribution followed the Eschmeyer's Catalog of Fishes (Fricke et al., 2023).As this study comprised only the Guanabara Bay, records obtained Keywords 1º search field "Guanabara bay" OR "Guanabara" OR "baía de Guanabara" OR "bahía de Guanabara" "AND" 2º search field "fish*" OR "teleost*" OR "elasmobran*" OR "pisces" OR "shark*" OR "ray*" OR "stingray*" OR "chondrichth*" OR "skate*" OR "bone fish*" OR "agnatha*" OR "osteichthy*" OR "actinopter*" OR "peixe*" OR "pesca*" OR "elasmobrânqui*" OR "tubar*" OR "raia*" OR "arraia*" OR "condrict*" OR "agnat*" OR "osteíct*" OR "pez" OR "tiburón" OR "tiburones" OR "raya*" A baseline of Guanabara Bay's ichthyofauna from local watershed rivers were not considered.Records from ichthyoplankton studies were also not included in our compilation.The historical baseline was built on data available until 2020, therefore we did not include records made after this year.However, we added the information of a few records made after this time-period in the Results section due to their ecological relevance.These include the first record of Galeocerdo cuvier (Perón & Lesueur, 1822) in Guanabara Bay and records of species that had not been recorded in the last decade (Bagre bagre (Linnaeus, 1766)) and Rhizoprionodon lalandii (Valenciennes, 1839)).
The selected records were used to build a baseline containing the currently valid name of the species, the year of the record, and the quadrant or quadrants where the species was recorded, if the information was available.When the exact year of collection was not indicated, the year of publication of the reference was considered as the date of the record.To generate a more complete inventory, the FishBase platform (https:// www.fishbase.se)and specific literature on each species were used to obtain information on (i) feeding and functional guilds in the estuary (standardized according to Elliott et al., 2007) and (ii) habitat (standardized according to Silva-Junior et al., 2016).Finally, information on the extinction risk of each species was considered at both the global and Brazilian level, according to the IUCN Red List of Threatened Species (http:// www.iucnredlist.org)and the Livro vermelho da fauna brasileira ameaçada de extinção (Subirá et al., 2018), respectively.Data analysis.One of the main difficulties of studies that aim to assess the species richness of a given locality is to determine whether the employed sampling effort was sufficient to accurately estimate the richness (Schilling et al., 2012).As our study consisted on building a reference database using previously generated data, the number of sources consulted was considered as a unit of sampling effort.Thus, four absolute richness accumulation (S) curves by effort were constructed using the R software version 3.6.0,considering one for the bay as a whole and one for each of the three compartments of its estuary (low, medium and upper).In all cases, the random sample-based rarefaction method was used (Gotelli, Colwell, 2001) employing 100,000 permutations.In order to further understand the results of the curves, the non-parametric estimator for incidence data Chao2 was applied to each curve (Chao et al., 2009) which, in addition to allowing the verification of curve stabilization (reaching an asymptote), also provides a series of other information (Tab.2).One of the advantages of using this estimator is the possibility to obtain "mg" values, since "g" values can be converted into percentages.In this context, if for a given "g" value extra collections are not necessary (null mg), then it is confirmed that the study in question was able to record the "g" of the percentage of total richness.A graph was also constructed for each rarefaction curve, where the x axis corresponds to "g" values and the y axis, to "mg" values.
Concerning the spatial richness distribution, accumulation of absolute richness (S) values of each water quadrant was plotted on the bay map employing the QGIS software version 3.16.5.As each quadrant has its own water surface area (discounting portions of land, such as islands and coastlines), species density (S/water surface area) was also calculated in each one of them to obtain comparable results.
Finally, considering the bay's history of environmental degradation and fishing exploitation, we expected the ichthyofaunal composition to change over the years.In this context, the temporal range of our baseline (1889 to 2020) was divided into decades to identify species that no longer occur in the bay, or that are at least rare now.
We considered recent all records made from 2010 to 2020, because since 2010 there were no one-off impact events (e.g., oil spills) that may have affected the ichthyofaunal composition.Therefore, for this study purposes, the 10 years period between 2010 and 2020 (last decade) represent the recent state of the bay.

RESULTS
Ichthyofauna richness and composition.The scientometric analysis resulted in a total of 176 published articles, 70 of which fitted the criteria described in Data compilation and were included in this study.Assembling all data compilation strategies, we considered a total of 84 different data sources.A total of 220 species (203 teleosts and 17 elasmobranchs) were recorded, distributed in 149 genera (136 teleosts and 13 elasmobranchs) and 72 families (61 teleosts and 11 elasmobranchs) (Tab.3).Regarding the Teleostei, a very asymmetrical richness distribution was noted among families given that 14 families make up about 50% of the total recorded richness.Among these, the Sciaenidae included the highest number of species (23), followed by Carangidae (14) and Haemulidae (8).Concerning elasmobranchs, the numerical variation of species between families was lower, with the Dasyatidae and Carcharhinidae including three species each, followed by Sphyrnidae and Rhinobatidae with two species.Other families of the Elasmobranchii are represented by just one species each.
Demersal species represented about 80% of the richness, distributed throughout species that inhabit soft substrates (58%), hard substrates (17%) or both (5%), while 44 species are classified as pelagic.These results are reflected in the feeding guilds identified, with 55% (120) of the species being considered zoobenthivores.The other categories have much lower values, with 39 opportunistic species, 21 zooplanktiovorous, 18 piscivorous, 14 omnivorous, four herbivorous and four detritivorous.
Elasmobranchs are vertebrates with a conservative life history (e.g., low fecundity, late sexual maturation, slow growth, high longevity, long gestational periods) and, therefore, have low replacement potential in the event of mortality from unnatural 18/34 ni.bio.br| scielo.br/niA baseline of Guanabara Bay's ichthyofauna causes (e.g., Hoenig, Gruber, 1990).Thus, it is not surprising that this group has a higher number of threatened species when compared to teleosts, a group with species generally presenting shorter life cycles and high population densities (e.g., Pratt Jr. et al., 1990).Among the ray and shark species recorded in this study, 77% are threatened globally (Vulnerable, Endangered or Critically Endangered), in addition to 23% considered as Near Threatened.Among teleosts, only 7% are threatened or Near Threatened globally, with 82% considered as Least Concern, 3% as Data Deficient, and 8% as Not Evaluated.A similar scenario was found at the Brazilian level, with 79% of the teleosts classified as Least Concern, 4% as Threatened, 5% as Near Threatened, 8% as Data Deficient, and 4% as Not Evaluated.As for the elasmobranchs recorded in Guanabara Bay at the Brazilian level, 47% of the species are assessed as Data Deficient, 35% are Threatened, 12% are Near Threatened, and only 6% are classified as Least Concern.
The rarefaction curve calculated for the estuary as a whole did not reach an asymptote, indicating that Guanabara Bay has an even richer baseline concerning fish species (Fig. 2A).In fact, the analysis by the Chao2 method estimated a richness of 249 species, 29 more than that recorded herein.However, about 88% of the ichthyofauna was successfully inventoried (Fig. 3A).The probability of obtaining a new species record if one more source was consulted would be only 0.046, while a significant effort would be required (319 new sources) for 100% of the fish species in the bay to be fully inventoried (Tab.4).Regarding only the lower estuary compartment, the results obtained were similar, with no stabilization of the rarefaction curve (Fig. 2B).However, in this case, the number of recorded species (188) was closer to the estimated (approximately 203 species), at 92% of the total ichthyofauna (Fig. 3B).In addition, the "q0" value was even lower (0.03) and the effort to obtain the totality of the lower estuary ichthyofauna would, again, be excessively high (m = 223.5).
Contrary to the lower estuary compartment and the Guanabara Bay as a whole, stabilization of the rarefaction curves was obtained for the middle and upper estuary compartments (Figs.2C, D), indicating that the records are sufficient to represent the species richness of these two portions of the bay.The observed and estimated richness values were very close in both cases and q0 values were less than 0.01 (Tab.4).Even though the "m" values were not null, they indicated a sampling of over 99% for these two compartments (Figs.3C, D).However, a new species was recorded in the upper estuary after 2020.On June 22, 2022, gillnet artisanal fisherman captured one specimen of the tiger-shark Galeocerdo cuvier on quadrant C4.This is the first record of the species in Guanabara Bay.The specimen was a juvenile female with total length of 1,80 m and its jaw is deposited in the MNRJ (MNRJ 53604).A baseline of Guanabara Bay's ichthyofauna Spatial distribution.In general, the region of the bay closer to the mouth of the estuary presented the highest values of both absolute richness (S) and species density (SD).Quadrants D7 and E7 at the entrance of the estuary were the richest and densest (Figs. 1, 4).Although D7 presented the highest number of species ( 127), E7 has the highest density due to its smaller water surface area.High S and SD values were also noted in the other lower estuary quadrants, with D4 having the lowest richness (43 species).However, D6 results were lower than expected (S = 55, SD = 2.34 sp./km 2 ) considering it is a transition region between D7 and D5, both of which are richer.
A high variation in S was observed in the middle estuary compartment.For instance, quadrants F4 and F5 presented less than 10 species, while quadrants C5 and E5 had over 60 species (Figs. 1, 4A).However, the quadrants of this compartment have very different water surface areas, making SD a more reliable measure for comparison.Even though quadrants B5, F5 and E6 have relatively small S values (18, 2 and 14, respectively), their small water surface area result in SD values above three (Figs. 1, 4B).Therefore, only quadrants C6 and F4 stand out with relativity lower densities when compared to the other quadrants of the middle estuary.
The upper estuary presented most quadrants with relatively lower values of S, with three quadrants with less than 10 species (D2, D3 and E2) and five with less than 20 species (B3, B4, C4, F2 and F3).A similar pattern was recovered for species density, with the upper estuary comprising the only portion of the bay with quadrants with SD values lower than one species per km 2 (B4, D3, E2, F2 and F3).All quadrants in the upper estuary compartment, except for C3 and E3, presented SD values lower than two species per km 2 .Indeed, quadrants C3 (S = 62, SD = 2.98 sp./km 2 ) and E3 (S = 73, SD = 3.07 sp./km 2 ) stood out in terms of richness, with S and SD values more similar to the ones recovered for the lower and middle estuary compartments (Figs. 1, 4).Temporal changes.From the 220 species considered, 84 were not recorded in the last decade (between 2010 and 2020) (Tab.5).Among these, 65 species were last recorded between 2000 and 2009, comprising 60 teleosts and five elasmobranchs.However, three of these elasmobranchs (Sphyrna tiburo (Linnaeus, 1758), S. zygaena (Linnaeus, 1758), and Pristis pristis (Linnaeus, 1758)) may have been recorded a long time before this timeframe.That is because their records come from past literature and ichthyological collections data presented at Buckup et al. (2000) who did not present the years that the records were made.Therefore, we considered the year of publication (2000) as the record data of those species.
Four species were recorded between 1960 and 1969 and three others were mentioned only between 1960 and1969 (Tab. 5).Last records of some species are considerably older, for instance Mugil curvidens Valenciennes, 1836, recorded in 1944, Narcine brasiliensis (Olfers, 1831), in 1938, and Parablennius pilicornis (Cuvier, 1829)    Nevertheless, fishing landing monitoring is still being carried out at Guanabara Bay, resulting in new records.After 2020, two new records of species shown on Tab. 5 were registered.On August 9, 2022, a 30 cm (total length) female Rhizoprionodon lalandii was captured by fishing at quadrant D2 and was deposited at MNRJ (MNRJ 53605).The species was recorded at the bay only once before in 1997.The other species is the teleost Bagre bagre previously recorded in 2005.The new record occurred on November 3, 2022, at quadrant F2.

DISCUSSION
The use of different sources for the compilation of past data was an efficient way to build a baseline of fish species from Guanabara Bay, as the different sources filled different gaps regarding the ichthyofauna survey.While the published literature provided more recent records, specimens deposited in ichthyological collections revealed more ancient occurrences, some dating back to the 19 th century.Scientific sampling and taxonomic monitoring of fish landings, in turn, revealed 13 species not reported by any other type of source.Since this survey is based on past records, it is important to consider the possibility of misidentification of specimens in the sources consulted.For the scientific sampling and fish landings monitoring this problem was likely minimized, since they were carried out by BioTecPesca/UFRJ and all specimens were identified by a specialist.The use of only published data also increases the reliability of the baseline, since all the articles used were peer reviewed by specialists.Other important measure was the exclusion of doubtful records of species that are not confirmed to occur in the state of Rio de Janeiro.Finally, our survey recovered a considerable level of internal data consistency, with the same species recorded in the same areas by different sources, increasing the reliability of the occurrence of these species.
Despite advances, some taxonomic questions still hinder the establishment of a more comprehensive list of fish species in the Guanabara Bay.For instance, Elops saurus Linnaeus, 1766 was considered the only species of the genus Elops in the western Atlantic before the description of Elops smithi McBride, Rocha, Ruiz-Carus & Bowen, 2010.However, the two species are anatomically similar, such that sympatry of the two species in the region cannot be ruled out at the moment.A similar situation refers to the distribution of Scomber japonicus Houttuyn, 1782. Fricke et al. (2023) indicates that its distribution is restricted to the Pacific Ocean.However, other studies have recognized S. japonicus as occurring in the southwestern Atlantic (Roldán et al., 2000;Perrotta et al., 2005) and specifically of Rio de Janeiro State (Alves et al., 2003;Menezes et al., 2003).Further studies are required to clarify the distribution of those species.
The total species richness recorded in the Guanabara Bay is considerably higher in relation to other tropical estuaries (Tab.6).The coast of the state of Rio de Janeiro is the richest portion in the Brazilian coast concerning estuarine fish species (Vilar et al., 2017), and Guanabara Bay stands out when compared to two other estuaries previously inventoried in the state (Sepetiba Bay and Mambucaba Estuary), having practically twice the number of species.Even though the large size of the Guanabara Bay contributes to a naturally greater richness, this factor alone is not able to explain the observed discrepancies.Sepetiba Bay, for instance, is similar in size to Guanabara Bay, but has practically half the number of species.Other example is the Bay of Malaga, in Colombia, that despite being much smaller (126 km 2 ) still presents three families, 36 genera and 17 species more than what we recorded at Guanabara Bay.The relatively high value of species richness in the Guanabara Bay is likely promoted by the diversity of environments and microhabitats, as the bay encompasses islands, mangroves, rocky shores, sandy beaches, artificial substates and muddy bottoms.In addition, the bay presents a wide variation of environmental conditions and gradients of salinity and nutrient distribution that are characteristic of estuaries (Vianna et al., 2012;Silva-Junior et al., 2016;Wolanski, Elliott, 2016).These conditions promote a wide variety of ecological opportunities, reducing competition and favoring the coexistence of a high number of species (Bello et al., 2012;Dolbeth et al., 2016).Furthermore, the bay conditions are seasonally influenced by a low intensity upwelling event.During spring and summer (November to March) changes in winds promote the outcrop of cold waters from the SACW mass, causing parts of the estuary to present subtropical temperatures (between 10 and 20 ºC) (Silva-Junior et al., 2016).This phenomenon allows species that only occur in deeper areas of the continental shelf to enter the estuary.
Concerning the absolute richness accumulation curves calculated, the upper and middle estuary curves stabilized, indicating that richness values recorded are close to the estimated value of those compartments.However, the record of Galeocerdo cuvier made in the upper estuary compartment in 2022 indicates that occasional species may occur even in the inner parts of the bay, especially when it comes to opportunistic highly mobile taxa.
Stabilization of the accumulation curves were not observed for the Guanabara Bay as a whole and for the lower estuary, both of which are likely to have larger values of richness than the ones recorded here.The lower estuary seems to have strongly influenced this result.Sampling effort required to reach an asymptote can be prohibitively large for environments with a high number of rare species (Chao et al., 2009).As the region closest to the adjacent coastal zone, the lower estuary is affected by the continuous inflow of oceanic water and is visited by many occasional opportunistic marine species, which functionally act as rare species.For instance, species associated with rocky shores from some beaches around the Guanabara Bay (e.g., Rodrigues-Barreto et al., 2017) probably enter the estuary during tide variations.However, their record can be hampered by the high turbidity that hinders visual census attempts.Indeed, the Chao2 index calculated unusually high "m" values for the lower estuary, indicating that approximately 223 new sources would be needed for the entire ichthyofauna to be inventoried in that compartment.Therefore, the non-stabilization of the lower estuary compartment may be preventing the stabilization of the Guanabara Bay's richness accumulation curve.This is a common situation in tropical environments, where different ecosystems have been sampled for decades without reaching an asymptote in the species richness (e.g., Gotelli, Colwell, 2011).
The distribution of the estuarine ichthyofauna is influenced by the interaction between coastal currents and the water from the local drainage basin, as well as by the degree of tolerance of each species to the salinity gradient (Camargo, Issac, 2003;Silva-Junior et al., 2016).Other important factors are the colonization capacity of different fish populations and the variety of habitats and biotic interactions that maximize interspecific coexistence (Bello et al., 2012;Dolbeth et al., 2016).The lower estuary is the compartment of the bay most influenced by coastal oceanic waters, allowing marine species to enter the compartment to feed (Nybakken, Bertness, 2005).These occasional marine species are likely to promote the high S and SD values in the lower estuary.The proximity to the coastal environments seems to produce a gradient in this compartment, with the innermost quadrants (D4 and E4) presenting lower values of S and SD than the outmost quadrants (D5, D7 and E7) (Fig. 1).The only quadrant that deviates from this pattern is D6.However, this is probably due to the difficulty of performing biological samplings, since this quadrant presents depths up to 50 m (Meniconi et al., 2012) and undergoes intense boat traffic.
The middle estuary is a transition area, presenting distinct spatial and temporal features.This compartment can be split into two portions that respond differently to the dry and rainy seasons, namely (A) the quadrants to the left of the central channel (B5, C5 and C6) and (B) the quadrants to the right of the central channel (E5, E6, F4 and F5) (Fig. 1).In general, the water column conditions during the rainy season are more variable than in the dry season, but in A the greatest amplitudes are related to temperature (minimum of 17 ºC and maximum of 28 ºC), while in B salinity is more variable (minimum of 18.8 S and maximum of 33.6) (Silva-Junior et al., 2016).The SD value of quadrant C6 differs from the rest of the quadrants of A, being the only one with SD lower than 3.0.As salinity does not vary considerably within this group, this low value is probably related to the fact that this location has been the subject of few studies and is not a BioTecPesca collection point, resulting in a lower sampling of this quadrant.In B, the F4 quadrant presented SD values lower than the rest (1.63 sp/km 2 ).In this case, lower SD values are probably caused by both a methodological factor (all records come from source 63) and to the innermost position of this quadrant, which makes it difficult for species that do not support lower salinities to inhabit.
Lower salinities may also play a role on the low S and SD values recorded in the upper estuary compartment, preventing marine coastal species from accessing the innermost part of the bay (Nybakken, Bertness, 2005;Silva-Junior et al., 2016;Souza, Vianna, 2022).Another important factor is that the upper estuary is the most environmentally impacted portion of the bay.In addition to receiving pollutants from rivers (e.g., untreated sewage and inorganic pollutants), natural water renewal is slow, resulting in low environmental quality anoxic zones.(Fistarol et al., 2015).However, two quadrants (E3 and C3) stood out with much higher S and SD values than the rest of the upper estuary.The absolute richness of 73 species found in E3 may be a result of the influence of the central channel during drought periods, when more saline waters advance to more inland estuarine regions, and the wide variety of habitats associated with the Paquetá island located in that region.Quadrant C3, on the other hand, is relatively far from the central channel and has no islands.The high S and SD values of C3 are thereby more likely to be related to collection efforts, as C3 was a BioTecPesca collection point.It also might be noteworthy that, both quadrants with higher S and SD values are within the influence of Protected Areas (Fig. 1).Quadrant E3 is very close to the APA Guapimirim, while the Barão de Mauá Municipal Natural Park is located on C3's coast.These Protected Areas may be relieving local anthropic impacts such as over-fishing, allowing for a higher number of species to be recorded in its vicinities.
Our results also suggest that temporal changes in the composition of Guanabara Bay's ichthyofauna occurred over the decades.As much as the absence of records between 2010 and 2020 is indicative of local extinction or at least of abundance reduction of species, more studies are still required to confirm this situation as shown by the recent records of Rhizoprionodon lalandii and Bagre bagre in 2022.Among the 84 species unrecorded from 28/34 ni.bio.br| scielo.br/niA baseline of Guanabara Bay's ichthyofauna 2010 to 2020 (Tab.5), 81% are not estuarine-dependent.Thus, the lack of more recent records for this non-dependent species may be due to their non-obligatory relation with the estuary coupled with declining environmental conditions.Another factor to be considered is the habitat of those species since there is a lack of recent studies in more consolidated substrate regions in the interior of the bay.
The decreased sampling efforts by BioTecPesca is especially important when it comes to the species last recorded between 1990 and 2010.The main sampling method used by BioTecPesca was bottom trawling, which results mostly in the capture of demersal species of unconsolidated substrate.Therefore, pelagic species like Hemicaranx amblyrhynchus (Cuvier, 1833), Rhinoptera bonasus, Rhizoprionodon lalandii, R. porosus, and Syngnathus pelagicus Linnaeus, 1758 would not be easily captured, so that their local extinction cannot be attested.This hypothesis is again supported by the new record of R. lalandii since this species was previously only recorded in 1997 but was captured by fisherman in 2022.Furthermore, collections made by BioTecPesca in 2013 and 2014 were less numerous and covered fewer locations when compared to the period between 2005 to 2007, thus hampering more recent records of species that do not commonly appear in scientific papers and are not recorded in fishing landings.
However, of the 70 species last recorded between 1990 and 2010, six are at risk of extinction at the global level (Mycteroperca microlepis (Goode & Bean, 1879), Paralichthys patagonicus Jordan, 1889, Pseudobatos percellens, Rhinoptera bonasus, Rhizoprionodon lalandii, and R. porosus), two are threatened at the national level (Hippocampus reidi Ginsburg, 1933 andHyporthodus nigritus (Holbrook, 1855)) and nine are threatened at both levels (Epinephelus itajara (Lichtenstein, 1822), E. marginatus, E. morio (Valenciennes, 1828), Hippocampus erectus Perry, 1810, Hyporthodus niveatus (Valenciennes, 1828), Pristis pristis, Pseudobatos horkelii, Sphyrna zygaena, and S. tiburo).For those species, abundance reduction at some level is likely to have occurred at Guanabara Bay.Considering the large number of studies and collections carried out after 1990, local extinction or great abundance reduction are also very likely to have occurred for the 13 species last recorded before 1990.These species span a wide functional range including pelagic and demersal species which are dependent or not dependent on the estuary.
Another worrying result recovered in this survey is that of the 17 elasmobranchs species, 10 (58.8%) were not recorded between 2010 and 2020.A well-documented case of elasmobranch regional extinction is the sawfish Pristis pristis, formerly occurring from northern Brazil to São Paulo and now restricted to the northern regions of Brazil (Fernandez-Carvalho et al., 2013).The disappearance of high trophic level predators is of concern for biodiversity conservation, as these animals play an important role in the ecosystem regulating prey populations.Besides, estuaries are extremely important for sharks and rays, serving both as a feeding area and as nursery grounds (Gonçalves-Silva, Vianna, 2018a;Plumlee et al., 2018).Signs of population reduction for these species in a large estuary such as Guanabara Bay may indicate the decline of elasmobranch populations throughout southeastern Brazil.
Even though Guanabara Bay still has a relatively rich ichthyofauna, with wide taxonomic and functional fish diversity, the implementation of management and conservation actions are paramount to reduce the loss of the biological richness recorded in our study in the more recent decades.There is a need for improvement of the environmental quality of the bay and adjacent regions.Many urban centers around .br | scielo.br/niA baseline of Guanabara Bay's ichthyofauna

FIGURE 1 |
FIGURE 1 | Guanabara Bay map, Rio de Janeiro, divided into five km x five km quadrants.Different shades of blue indicate which estuary compartment (upper, middle or lower) the quadrant belongs to.

FIGURE 2 |
FIGURE 2 | Rarefaction curves for Guanabara Bay's ichthyofauna richness, Rio de Janeiro, Brazil. A. For the bay as a whole, and for the estuary compartments separately: B. Lower estuary, C. Middle estuary and D. Upper estuary.

FIGURE 4 |
FIGURE 4 | Distribution of absolute richness (A) and species density (B) per km 2 at Guanabara Bay, Rio de Janeiro, Brazil.

TABLE 1 |
Keywords used in the scientometric search on fish at Guanabara Bay, Rio de Janeiro.

TABLE 2 |
Variables related to the non-parametric Chao2 estimator.The t, T, S obs, Q1 and Q2 values are used to calculate S est, q0, m and mg.

TABLE 3 |
Species reported at Guanabara Bay and the sources, record dates and quadrants (column Q) in which these records occurred.

TABLE 4 |
Chao2 parametric estimator values for Guanabara Bay, Rio de Janeiro, Brazil, as a whole, and for the lower, middle and upper estuaries separately (t = 84).

last recorded Dates Species last recorded from 2000 to 2009
, in 1915.Data of those records are again based on voucher specimens at MNRJ.

TABLE 5 |
Species not recorded after 2010 at Guanabara Bay, Rio de Janeiro, Brazil, until 2020.

TABLE 6 |
Absolute ichthyofauna richness in tropical estuaries in Brazil and worldwide according to the available literature.