A decade on: an updated assessment of the status of marine non-indigenous species in Brazil

In order to prevent the introduction, control or eradicate non-indigenous species (NIS) which threaten native species, habitats or ecosystems, an essential first step is that countries have and keep up-to-date lists of species non-indigenous to the region. The last list of marine NIS for Brazil was published a decade ago. We compiled an updated list of marine NIS and revised the species’ statuses. Onehundred-thirty-eight marine species in 15 classes or functional groups are NIS in Brazil. Brazilian states with greater maritime commerce (greater market share) had more marine NIS. From the period between the 1950s to 2010, the number of NIS at least doubled each decade. We identified a significant mismatch (underestimation) between the previous list and this study, which seemed to be due to improved scientific knowledge and an often considerable lag between first record (detection), identification and communication of new marine NIS. Currently Brazil has 19 invasive, 76 established and 43 detected marine NIS, an increase of 160% in ten years compared to the previous national list. We recommend that Brazil implements a national database as a rapid, appropriate, flexible and cost effective method of monitoring trends in NIS introductions.


Introduction
The Convention on Biological Diversity (to which Brazil is a party, Brasil 1998) states that consenting parties should prevent the introduction, control or eradicate those alien species which threaten ecosystems, habitats or native species. In Brazil this convention was ratified in 1994, with the Ministry of the Environment being the principal entity responsible for planning and decision making with regard to non-indigenous species (NIS) and biological invasion ). An essential first step is that countries have and keep up-to-date lists of NIS recorded for their region. According to Sliwa et al. (2009) once lists are adopted into the 3 1 management framework, they are often used for long periods without revision. However it is important that these lists are constantly reviewed and updated to incorporate new introductions and changes in species status based on further research (Sliwa et al. 2009). This information may be used to assess management initiatives and targets.
The Brazilian coastline extends for about 8000 km from Cape Orange (4°N) to Chui (34°S) comprising a range of tropical and sub-tropical ecosystems such as sandy beaches, mangrove forests, coral reefs, rocky shores, coastal lagoons, saltmarshes, oceanic islands and sandy to muddy bottoms from the infralittoral realm down to the shelf break (Ferreira et al. 2009). Further offshore the Exclusive Economic Zone of 200 nautical miles and the proposed extension called the Amazônia Azul, or "Blue Amazon" (Wiesebron 2013) form a very large oceanic biome. According to Ferreira et al. (2009) this makes the country a major receptor and donor of tropical and subtropical organisms amongst the world's oceans.
The INFORME/PROBIO Project, which was conducted in 2005, provided the first comprehensive list of non-indigenous species in coastal areas of Brazil (Ferreira et al. 2009). The I National Report on Invasive Exotic Species affecting the marine environment (fauna, flora and microorganisms) (I Informe Nacional sobre as Espécies Exóticas Invasoras que afetam o ambiente Marinho (fauna, flora e microrganismos)) resulting from the study was published in 2009 by Lopes et al. At the time of that study 58 marine NIS were known from Brazil ); the species that made the most impact were reviewed by Ferreira et al. (2009). In this paper, we compile, update, systematize and synthesize currently available information on marine NIS in Brazil in order to characterize the frequency of introduction events as well as reassess the status of the species a decade on from the INFORME/PROBIO list published in 2009 by Lopes et al.

Materials and methods
Based on the INFORME/PROBIO list and data provided therein  We classified the status of each species by using the same criteria and categories as those described by Lopes and Villac (2009), as these have been adopted by the Brazilian Federal Government. Lopes and Villac (2009) recognized a status hierarchy as the following: Contained: when the presence of the NIS has only been detected in controlled artificial environments, totally or partially isolated from the natural environment (e.g. commercial aquarium, cultivation for scientific purposes, vessel ballast water tanks, etc.); Detected: when the presence of the NIS was detected in the natural environment, but with no subsequent increase in its abundance and/or its dispersion (considering the time horizon of research or surveys); or, alternatively, without further information on the population situation of the species (e.g. an isolated record); Established: when the NIS has been detected on a recurrent basis, with its full life cycle in nature and evidence of population increase over time in a restricted or broad region, but without apparent ecological or socioeconomic impacts; Invasive: when the NIS has an abundance or geographic dispersion that interfere with the survival capacity of other species in a broad geographic region or even in a specific area (Elliott 2003), or when the established species causes measurable impacts on socioeconomic activities or human health.
When a species had been assigned different statuses in different studies, we used the classification criteria based on the most recent study available (e.g., compare Farrapeira et al. 2011 andRocha et al. 2013). Species native to Brazil which demonstrated human-mediated range shift (sensu Lonhart 2009) were also included. Furthermore, we noted the occurrence by Brazilian state and the year of first record in the country. This was taken as the year of collection, or record, if reported in the reference(s), or the year of the reference publication when the year of collection or record was not provided. The nomenclature was updated in accordance with the World Register of Marine Species (Worms Editorial Board 2019). Contained or cryptogenic species were not included in the present list of NIS. In order to compare the status of different taxonomic or functional groups, we organized records by functional group, class, order and family. In order to describe temporal changes, we calculated cumulative records and estimated current detection rates. We also compared the status according to  with the revised status noted here. In order to describe spatial patterns, we also documented species by Brazilian state. As states are of different sizes, we also converted these data to a per km basis using coastline length data obtained from https://pt.wikipedia.org/wiki/Litoral_ do_Brasil accessed on 26/12/2018 (Supplementary material Table S1). To test whether the volume of shipping explained the difference in the abundance of NIS between states, we used a simple linear regression between the number of NIS per state and the "Market share" which is the proportion of the national tonnage of shipping (imports + exports) for which each state is responsible (ANTAQ 2018; Table S1).  . Zooplankton category includes all planktonic Hexanauplia, Scyphozoa and Hydrozoa; * = zoobenthic species of these classes.

Results
We compiled information from more than 120 data sources (Table S2). We reviewed all studies and identified and assigned status to 138 marine NIS recorded from Brazil up to May 2019. Contained or cryptogenic species were not included in the list. The updated summary and status list is presented in Table S3. The NIS comprised 15 classes or functional groups ( Figure 1); Malacostraca was the class with most NIS (n = 26), followed by Ascidiacea (n = 22), Polychaeta and Gymnolaemata (n = 12 each). Elasmobranchii (n = 1), Ophiuroidea and Gastropoda (n = 2 each) and phytoplankton (n = 3) were the groups with the least NIS. The group with the largest number of invasive (sensu Methods) NIS was the Anthozoa (n = 7), followed by the phytoplankton and Bivalvia (n = 3 each) (Tables S2 and S3).

Distribution in space
Up to May 2019 at least three NIS have been documented in each of the 17 Brazilian coastal states, but most were concentrated in southeast Brazil. The state with the greatest absolute number of NIS was São Paulo with 73 species (52.9% of total), followed by Rio de Janeiro with 69 (50%), Bahia with 45 (32.6%) and Paraná with 43 (31.2%); the state with the lowest number  Table S2). However, considering and correcting for the significant variation in coastline extension between states on a per km basis (Table S1), there was a trend of increasing numbers of NIS from north to south to Paraná, which although a small state had the highest number of NIS per 100 km of coastline ( Figure 3). The states of Piauí and Pernambuco also had a higher number of NIS than expected by the north-south trend ( Figure 3).
The most widespread NIS in Brazil is currently the Pacific White Shrimp Penaeus vannamei (15 states; Table S2). Other widespread NIS are the Indo-Pacific swimming crab Charybdis (Charybdis) hellerii which occurs in 11 states, the copepod Temora turbinata and purse oyster Isognomon bicolor in 10 states, the hydroid Cordylophora caspia, snowflake coral Carijoa riisei and mudsleeper Butis koilomatodon in eight states, and the barnacles Amphibalanus reticulatus and Megabalanus coccopoma and sun corals Tubastraea coccinea and Tubastraea tagusensis in seven states each (Table S2). Only one species, the macroalga Caulerpa scalpelliformis, has a demonstrated human-mediated range shift as its natural southern limit of distribution in Brazil is Espírito Santo state but is invasive in the state of Rio de Janeiro. There were several NIS previously reported in Brazil in 2009 which expanded their ranges into other states over the last 10 years. These species included the mudsleeper Butis koilomatodon (into seven additional states), the orange sun coral Tubastraea coccinea (six), yellow sun coral Tubastraea tagusensis and the muzzled blenny Omobranchus punctatus (four each) and the boring bivalve Leiosolenus aristatus (three) ( Table S2). The alga Anotrichium yagii, mussel Mytilopsis leucophaeata, sabellid polychaete Branchiomma luctuosum, portunid crab Scylla serrata and the xanthid crab Pilumnoides perlatus also expanded their ranges (one additional state each).
The total number of NIS per 100 km of coastline in each state was dependent on the volume of market share (exports + imports, Table S1); a significant though not a very strong relationship was observed (NIS per 100 km coastline = 1.5748 × Market share + 14.383, Adjusted R² = 0.2494, p = 0.01; Figure 4). However when two outlier states which have exceptionally large exports of specific commodities were removed (Espírito Santo and Maranhão), a stronger relationship was observed (NIS per 100 km coastline = 3.466 × Market share + 9.918; n = 14, Adjusted R² = 0.71, p ≤ 0.001; Figure 4).

Change over time
The first record of a non-indigenous marine species in Brazil was from 1860 (the bryozoan Amathia verticillata) (Miranda et al. 2018) followed by the ascidian Styela plicata recorded in Rio de Janeiro in 1883   (Figure 5 and Table S2). The third record occurred 40 years later, in 1925 (the ascidian Clavelina oblonga) (Rocha et al. 2012). Up through the 1960s additional NIS were sporadically reported, but from 1949 onwards new records of NIS have at least doubled (≥ 100% increase) every decade up to a total of 99 in 2008 ( Figure 5). Over the last decade the number of new records of NIS has increased to 138 which represents a 25% decadal increase according to our dataset. The only group which did not increase in NIS abundance over the last decade was the phytoplankton (Figure 1). The state which exhibited the most substantial increase in NIS over the last decade was Paraná (25.5 species/100 km), then Pernambuco (8.6 species/ 100 km) while Rio Grande do Sul, Amapá, Pará and Maranhão only increased by < 1 species/100 km (Figure 3).
The current status of marine NIS and the change in status over the last decade as compared to  is summarized in Table S3. As some  (Tables S2, S3). According to this comparison, the number of NIS in Brazil has increased by 160% over the last 10 years. There has been no loss of any marine NIS in Brazil over the last 10 years, except for species that previously had been described as non-indigenous but that are currently considered as cryptogenic or native. Compared to the previous list, there was a 347% increase in the number of established species and a 111% increase in invasive species. The lowest increase was in the number of detected species (59%). Five species previously considered established were considered invasive, and eleven species that were previously considered to be detected were considered established (Table S3).

Discussion
Ten years on from the first national effort to compile a list and define the status of marine species non-indigenous to Brazil, published by the Ministry of the Environment ), much has changed. The effort by    (Table S2). In 2013 the National Biodiversity Targets established that the National Strategy on Invasive Alien Species should be fully implemented in Brazil by 2020, including the participation and commitment of the states in the formulation of a National Policy. Central to the national strategy is the need to identify and list NIS. The present study clearly demonstrates the need for regularly updating national lists of NIS.
This study is a data compilation, and therefore we compiled previously published information and assigned status according to previously established classification criteria. This must be done in an accurate and responsible way but even so the literature contains controversy regarding NIS statuses. Of the 58 NIS listed by , eight were not included in the present study. Five of these are now considered cryptogenic and include the sponge Paraleucilla magna Klautau, Monteiro andBorojevic, 2004 (Rocha et al. 2013;Cavalcanti et al. 2013) and the bryozoan species complex Schizoporella errata (Waters, 1878) (see Rocha et al. 2013;Miranda et al. 2018), both previously considered established, the bryozoans Licornia diadema (Busk, 1852) and Virididentula dentata (Lamouroux, 1816) (see Miranda et al. 2018) and the ascidian Bostrichobranchus digonas Abbott, 1951(Rocha et al. 2013, all previously considered detected. The isopod Sphaeroma annandalei Stebbing, 1911 was described as the new species Sphaeroma silvai Khalaji-Pirbalouty and Waegele, 2010, and thus lost its NIS status (Khalaji-Pirbalouty and Waegele 2010). The seaweed Kappaphycus alvarezii (Doty) Doty ex P.C. Silva, which was introduced intentionally into Brazil for farming, has not been able to establish itself naturally outside the culture systems (although broken, unfixed fragments are occasionally found on adjacent bottoms), so we considered this to currently be a contained species (Castelar et al. 2009;Lima et al. 2018).
With regard to the eighth species, the mussel Perna perna (Linnaeus, 1758), that was initially treated by  as a NIS, much has been discussed about the origin of this bivalve [see contrasting views of Cunha et al. (2014), Pierri et al. (2016) and Oliveira et al. (2017)]. Based on molecular, ecological and archaeological studies, the species may have been historically transported to Brazil from Africa on slave boats or be native. It is an edible species which is harvested and farmed and is of significant commercial importance; maybe for these reasons it has been termed "naturalized" by Oliveira et al. (2017). We did not include it in the present list as additional molecular, ecological and archaeological studies of different populations are required to better clarify the origin of the species. As it is of uncertain origin, it should be considered cryptogenic (a conservative category of obscure or unknown origin; Carlton 1996). The fact that a number of previously listed species (n = 6) have had their status redefined as cryptogenic attests to a more conservative scientific approach as invasion biology matures in Brazil (Carlton 1996).
Both morphological characteristics and molecular studies can aid in determining the invasive status of questionable species.  listed the bivalve Mytilopsis leucophaeata as a NIS in Pernambuco state but a subsequent molecular study by Fernandes et al. (2018) identified that population as M. cf sallei also a NIS for Brasil; however in the same study they also identified a new population in Rio de Janeiro state as M. leucophaeata, so the species maintains its status as NIS for Brazil. In addition to molecular studies, identifying morphological characters is important when determining NIS. The ascidian Ciona intestinalis, also known as a complex of two species, was later described in 2015 as C. intestinalis and C. robusta (Brunetti et al. 2015). Here we conservatively maintained the species as Ciona intestinalis until future studies establish which species actually occur in Brazil. Five new taxonomic classes not previously listed as NIS in Brazil were added to the list and include the following: two ophiuroids (Ophiactis savignyi and Ophiothela mirabilis), two gastropods (Bulbaeolidia alba and Eualetes tulipa), two Scyphozoa (Cassiopea andromeda and Phyllorhiza punctata), five hydrozoans (Podocoryna loyola, Cordylophora caspia, Blackfordia virginica, Cnidostoma fallax and Garveia franciscana) and one Elasmobranchii (the whitetip reef shark Triaenodon obesus).
The 11 species which were described as detected in 2009 and are currently considered established changed their classification because they have expanded their ranges and/or have been detected on multiple occasions. The five invasive species that were previously considered established had their status changed because they are now known to cause some type of damage, be it environmental, economic and/or to public health (Table S3). For example in 2009 the muzzled blenny Omobranchus punctatus was considered to be established and possibly an invasive species. Currently, this herbivorous fish, which scrapes algae from the bottom, has been described as competing with the native molly miller Scartella cristata (Linnaeus, 1758) and thus gained the status of an invasive species (Soares et al. 2011).
The increase in the number of established species and the decrease in detected species demonstrates that we are still ineffective in completely controlling the movement of NIS and that we must focus on management plans to stop NIS increasing their area of coverage and eventually becoming invasive (sensu Methods). Rather than focus exclusively on prevention of introduction, the ideal invasive species policy should also involve a combination of other strategies such as risk assessment, efforts to detect new species and activities to monitor and control existing populations of newly detected species. If an incipient population of a species is detected before establishment and spread, subsequent control costs may be reduced and eradication may be a viable option (Mehta et al. 2007). It is important to remember that so far in Brazil no established or invasive marine NIS has been eradicated after its introduction. That having been said, two species, the lionfish Pterois volitans and the octocoral Clavularia viridis, were both controlled on detection. They have not since been detected in Brazilian waters (Ferreira et al. 2015;Mantelatto et al. 2018) and may thus be considered eradicated.
Those species which have become widespread use multiple mechanisms of expansion. Some have been introduced multiple times for aquaculture (e.g., Pacific White Shrimp Penaeus vannamei, ) or on shipping associated with the offshore oil and gas industry (e.g., sun corals Tubastraea coccinea and Tubastraea tagusensis, and the associated purse oyster Isognomon bicolor; Creed et al. 2017). Others, such as the snowflake coral Carijoa riisei and the barnacles Amphibalanus reticulatus and Megabalanus coccopoma, are commonly found on artificial floating substrates (Ferreira et al. 2006;Carlton et al. 2011). The Indo-Pacific swimming crab Charybdis (Charybdis) hellerii, the copepod Temora turbinata and mudsleeper Butis koilomatodon were probably spread by currents, swimming and/or larval dispersal ).
In cases where species are introduced into new habitats for economic benefit or to meet developmental needs, there may be an initial economic gain, but if a species becomes invasive it can cause serious economic and ecological damage (Molnar et al. 2008). These risks of invasion are often not considered during the decision process of whether to introduce a species (Naylor et al. 2001). For example, ten years ago the Pacific oyster Crassostrea gigas was considered a contained species because it did not reproduce naturally on the Brazilian coast, and the cultivation of this species depended on the constant imports of seed or reproduction in the laboratory ). However C. gigas has now been found in natural banks on the south coast of Brazil, at a distance of up to 100 km south of the farms where they are cultivated (Melo et al. 2010). For this reason, C. gigas is now considered an established NIS.
Our study showed that the states with the greatest number of NIS are also those that have substantial movement of marine commerce. These ports usually also offer services for the off-shore oil industry, including rig harboring, and are important marine navigation hubs. These are three of the most important vectors of introduction of marine species (Ruiz and Carlton 2003;Silva and Souza 2004;Ferreira et al. 2006;Yeo et al. 2009). As outliers, we excluded two states from the analysis, Maranhão (MA) and Espírito Santo (ES), which both export very high tonnages of minerals from the interior of Brazil. We used market tonnage as a proxy for species dispersal as that data were available but frequency with which ports receive vessels and their origin was also important. Furthermore, some states (such as ES) are probably understudied while others (Rio de Janeiro -RJ, São Paulo -SP and Parana -PR) contribute to the bulk of previously published data (Figure 4). Identifying the spatial distribution and most common pathways for the introduction of harmful marine species can help to focus monitoring efforts, provide information for risk assessment and inform and support international policies to prevent such introductions (Molnar et al. 2008). In this respect it would be useful for management if future studies attempt to 1) identify the principal pathways and vectors of 4 1 introduction for these NIS, if this information is available; 2) identify which states are most critically threatened by NIS (by analyzing first introductions at the state level).
When we compared the timeline of reports of marine NIS in Brazil over the last 10 years to , we identified apparently incongruent results. For example we estimated a 160% increase in NIS over the last decade compared to Lopes et al.'s (2009) study but only a 25% decadal increase when comparing our own results. This was because by 2008 there were 99 marine NIS in Brazil compared to only 53 listed by  (see the point and line indicated in Figure 5). Bearing in mind that we used the year of first record (collection date, record date or publication date, in that priority order) for each new NIS, the reasons for this mismatch are various and could include the following: 1) while Lopes et al. was published in 2009, the data had been collected some time previously; 2) several NIS species included here, which were known in Brazil pre-2009, were not considered to be NIS at that time; 3) novel NIS may take some years to be formally identified, or even described, after first being recorded (but the date remains the same); 4) records of novel NIS may take years to get published. For example the brittlestar Ophiothela mirabilis was first detected in 2000 but the first published record was in 2012 by Hendler et al. Similarly, the octocoral Stragulum bicolor was first detected in 2002 in Brazil but only described (and as a new species) by Van Ofwegen and Haddad in 2011.
Given the above, it should be recognized that the apparent reduction in the rate of new records of marine NIS this last decade may also simply be an artifact of the lag between new NIS being detected and the community becoming aware through communication or publication of those records. Nevertheless, it may also be the case that the date of first record does not actually reflect the year of introduction, since many species may remain undetected for a while, which generally results in the opposite pattern of an apparent increase in the rate of introductions that is an artifact of increased awareness and sampling effort. These observations further confirm the need for NIS lists to be constantly reviewed and updated to incorporate newly detected NIS and changes in species status based on further research (Sliwa et al. 2009). Brazil has not yet implemented a national database of marine NIS which we consider may be the most appropriate and cost effective method of moving forward in updating and maintaining current the national NIS lists as part of the National Strategy on Invasive Alien Species.