Settlement plate surveys were conducted from 2000 to 2020 in 35 bays across 35 degrees of latitude (26°N to 61°N) in both estuarine and marine areas. Study sites included 8 embayments on the United States (U.S.) Atlantic coast, 1 on the Canadian Atlantic coast, 4 in the U.S. Gulf of Mexico, 21 on the U.S. Pacific coast, including 7 in Alaska, and 1 on the Canadian Pacific coast (Fig. 1, Table 1). Embayments included some of the largest shipping ports and encompass a wide range of climatic conditions from subtropical to subarctic. Standardized survey arrays of PVC settlement plates were deployed at 1 m below MLLW for three months during summer, to coincide with period of high reproduction and recruitment of many invertebrates (Simkanin et al. 2016; Jewett et al. 2022). All bays included marine habitat (defined as sites where salinities were ≥ 20 PSU at deployment and retrieval) as the primary focus. A few bays also supported estuarine (salinities 5–20 PSU) habitats; see Table 1 for details. Ten sites were sampled within each bay, and a minimum of 5 plates were examined from each site yielding at least 50 fouling plates analyzed per embayment. Bryozoans were identified to species level, when possible, while alive on plates, and then vouchered in 95% ethanol or 10% formalin and catalogued for further verification or identification by us, and a subset of the more difficult taxa were sent out for examination by bryozoan taxonomist colleagues. Since 2015, vouchered samples have routinely included both paired morphological and genetic samples, the latter to be reported elsewhere when sequence data is published. All vouchered specimens from this study are deposited at the Smithsonian Environmental Research Center (SERC) in Edgewater, Maryland.

Figure 1. 

Number of non-native species of bryozoans in the continental US and Canada by coast, bay, and latitude. For bay abbreviations see Table 2.

While all bays were sampled at least once in the twenty-year period, several bays were sampled in multiple years. Repeated surveys were conducted at selected bays in Alaska, California, Florida, and Virginia, as part of a long-term study of temporal changes across taxa (and are the focus of separate papers, in preparation). Here, we include records of non-native bryozoan species detected in these surveys.

Species were assigned native, introduced, or cryptogenic status based on the criteria of Chapman and Carlton (1994) and Carlton (1996), see Suppl. material 1. For the present analyses, we included only those species considered introduced and detailed cryptogenic and native species are in Suppl. material 1. Since many of the species identified are considered cosmopolitan and their origin is either unknown or uncertain, we utilized the framework of Darling and Carlton (2018) to categorize species in greater detail as 1) eucosmopolitan—naturally cosmopolitan, 2) neocosmopolitan—human mediated cosmopolitan, 3) pseudocosmopolitan—a cosmopolitan species by virtue of its uncertain taxonomic status of genetically distinct sublineages, and 4) unresolved cosmopolitan—for which we have little or no current information about a species status. A species might be classified as native on one coast and introduced on another based on these criteria.

For the purpose of our analyses, bryozoan species were considered established if, 1) two or more distinct colonies were found in a bay, 2) there were multiple records over multiple years for a location (for example 1 colony in year 1 and 1 colony in year 2 in San Francisco Bay would be considered established), or 3) the species was reported as established in the literature or through personal communications with colleagues. For certain species, the most recent records are 20 or more years old, based on our surveys, on published records (below), or both, and thus may not represent their current status at certain sites. We detected more non-native species on settlement plates than reported herein, but only as single records so we considered their establishment uncertain. These records include Biflustra cf. grandicella (Canu & Bassler, 1929), Biflustra irregulata (Liu, 1992) and Crisularia cucullata Busk, 1867; see Suppl. material 1 for localities.

To compile as complete a list as possible of fouling bryozoan species in the continental U.S. and Canada, we assembled data from three sources: 1) our standardized field surveys (above), 2) introduced species bioblitzes (n = 4, Sitka, Alaska 2010 and Ketchikan, Alaska 2012 (Ruiz et al. unpublished), Valdez, Alaska 2016 (Ruiz et al. 2017), Dutch Harbor, Alaska 2018, and rapid assessment surveys (RAS) (n = 9, including: east coast: New England 2010 (McIntyre et al. 2013), New England 2018 (Kennedy et al. 2020); New York and southern New England 2019 (Pederson et al. 2021), and west coast: San Francisco Bay, California 2004 (Cohen et al. 2005), southern California bays 2000 (Cohen et al. 2005), Los Angeles/Long Beach 2013–14 (Vilas et al. 2016), Los Angeles/Long Beach 2018 (Stolzenbach et al. 2021), Puget Sound, Washington 1998 (Cohen et al. 1998), Willapa Bay, Oregon 2000 (Cohen et al. 2001), and 3) a literature review of over 1000 publications and reports, including published papers, grey literature, online resources, and the National Estuarine and Marine Exotic Species Information System or NEMESIS (Fofonoff et al. 2018).

We examined the literature for species found in the U.S. and Canada, but if known distributions extended into Mexico, this information was also included in Table 2 and Suppl. material 1, including these unresolved taxa, although we did not exhaustively examine records beyond the U.S. southern border.

In reviewing the literature, we at times found records for species remarkably similar to our species’ records, which might in fact represent the same taxa, but for which we have used a different name. Given the taxonomic difficulties around these groups and our inability to examine the specimens reported in these papers, we have not used these unresolved species in our analyses as we view them as unverified. For example, Alcyonidium species were excluded from analyses due to the lack of taxonomic resolution and resultant difficulty in determining their native status (Ryland and Porter 2012). Locality data from plates, tentative species names and native range designations are given in Suppl. material 1.

Date of first record was assigned based on either the first mention of a non-native species in the literature or its appearance in our plate surveys, whichever came first (Table 2). If a publication did not specify a collection date, we used the date of the publication. We note that “first record” data reflect biases in sampling location, timing, and effort, and available taxonomic expertise, and these records may often include substantial publication lag times. Regardless, we emphasize that first record dates are unlikely to represent the actual date of introduction, which may have occurred years or decades earlier.

Where possible, Museum specimens were examined to determine date of first records, including examining Membranipora spp. and Conopeum spp. at the National Museum of Natural History, Washington, DC (Invertebrate Zoology Collections Database https://collections.nmnh.si.edu/search/iz/), the California Academy of Sciences, San Francisco, CA, the California Department of Fish and Wildlife collections housed at the Santa Barbara Museum of Natural History, Santa Barbara, CA, and Watersipora spp. collections in the Invertebrate Zoology Department of the Yale University Peabody Museum, New Haven CT (https://peabody.yale.edu/explore/collections/invertebrate-zoology).

We undertook the following analyses:

1. The total diversity of non-native fouling bryozoans across our sampled sites, and their geographic patterns (Table 1, 2, Figs 1–3).
2. The number of non-native bryozoans documented over time, to determine if repeated sampling efforts revealed new records.
3. Patterns of non-native bryozoans by life-history aspects including Order, growth form (upright or encrusting) and larval type (planktotrophic or lecithotrophic), the latter to examine potential dispersal capabilities (Suppl. material 1).
4. Transport vector(s), based on data from the literature. For all the species, multiple vectors were considered likely. Vectors included (1) Ballast Water (BW)—the ballast systems of ships, including entrainment of bryozoan larvae (such as those of membraniporines and some ctenostomes) and entrainment of fragments of seaweed, seagrass, wood, or plastic, which may support bryozoan colonies; (2) Ship Fouling (SF)—the hulls and other underwater surfaces of ships, and (3) Commercial oyster movements (CO)—accidental transfers with oyster transplants or equipment (Table 2).
5. Relative abundance and species richness of non-natives as a proportion of bryozoan communities across our site and ocean (Figs 1, 2).

We have excluded from our analyses those taxa that are not identified to species level. Yet those taxa unique to a site, even if unresolved at the species level, are included in Suppl. material 1. We have also included bryozoan species from freshwater sites (< 5PSU) in Suppl. material 1, but their sampling was limited to a few bays, so they are not included in the comparative analyses across bays or coasts.

Figure 2. 

Native range of bryozoans introduced to the US and Canada by ocean of introduction.

Figure 3. 

Number of non-native bryozoans as a function of latitude and ocean. The regression for the Pacific bays was run twice: with (solid red line) and without (dotted red line) the outliers or ‘hot spots’ circled in yellow. From left to right they include San Diego, Long Beach, and San Francisco, California, and Coos Bay, Oregon).

A generalized linear mixed model (GLMM) with negative binomial error distribution was conducted to model non-native species (NNS) richness as a function of latitude (nested within Ocean as a random factor) and latitude and ocean as fixed and crossed predictors. The lme4 package (Bates et al. 2015) was used to fit the GLMM, and model fit was checked using simulations constructed with the DHARMa package (Hartig et al. 2022). To further examine differences by bay and by ocean and to determine which species contributed to the patterns observed, we applied a PERMANOVA, a non-parametric permutational analysis of variance using distance-based matrices of resemblances among the bays (Anderson 2001). Latitude was utilized as a covariate, nested within ocean, and with permutations restricted within each ocean using the “strata” option. Multivariate analyses and a non-metric multidimensional scaling (nMDS) plot of dissimilarities were carried out using the vegan package (v2.6-4; Oksanen et al. 2022) in the R statistical computing environment (R Core Team 2023). Bray-Curtis similarities were employed, with 999 resamplings. All other graphical summaries and statistical comparisons were generated in Microsoft Excel.

Over 40,000 bryozoan specimens were collected and identified from nearly 4000 settlement plates. One hundred and seventy-eight bryozoan species were recorded over the 35 separate bays and estuaries. Forty-two non-native species were detected on our panels across all sites, representing 24% of total bryozoan species richness. Total non-native species richness for bryozoans per bay ranged from 0 to 30 across all coasts. The percentage of non-native species per bay ranged from 0% (sites in Alaska) to 72.73% (San Francisco Bay, CA) with a mean of 26.17% (se = 3.14%) and median of 22.22%.

In total, we documented 48 non-native species in our analyses, including 42 from our panel surveys and 6 from the literature and other surveys, including 13 Ctenostomata (27%) and 35 Cheilostomata (73%) (Table 2). Nine of these species are new records for the U.S. and Canada documented in our studies, including 3 newly reported herein: Amathia brasiliensis Busk, 1886, Nolella sawayai Marcus, 1938, and Celleporaria umbonatoidea (Liu & Li, 1987), two previously published online in reports to the State of California: Hippopodina iririkiensis Tilbrook, 1999 (Ruiz and Geller 2015) and Amathia citrina (Hincks, 1887) (Ruiz et al. 2023), and 4 previously published in McCann et al. 2007: Hippoporina indica Pillai, 1978, Arbopercula bengalensis (Stoliczka, 1869), Sinoflustra annae (Osburn, 1953) and Celleporaria pilaefera (Canu & Bassler, 1929).

Two species are newly reported from this work as non-native: Amathia dichotoma (Verrill, 1873) (previously published in Ruiz and Geller 2015), introduced to the Pacific coast, and Celleporaria brunnea (Hincks, 1884), introduced to the Gulf of Mexico. Twenty species previously recognized as non-native, are here documented with one or more new locality records, including 7 species under now-corrected or revised names: these include Amathia brasiliensis (formerly identified as A. distans Busk, 1886), Watersipora subatra (Ortmann, 1890) (formerly identified as W. subtorquata (d’Orbigny, 1852)), Bugulina cf. foliolata (formerly identified as B. flabellata (Thompson in Gray, 1848) or B. fulva Ryland, 1960 on the west coast, Schizoporella variabilis (formerly identified as S. errata (Waters, 1878)), Schizoporella japonica (Ortmann, 1890) (formerly identified as S. unicornis (Johnston in Wood, 1844)), Tricellaria inopinata d’Hondt & Occhipinti Ambrogi, 1985 (formerly identified as T. occidentalis catalinensis (Robertson, 1905) or T. occidentalis (Task, 1857), and Watersipora subtorquata (formerly under several names, including W. cucullata (Busk, 1854) and W. subovoidea (d’Orbigny, 1852)) (Table 2, Suppl. material 1).

Based on data from both our plates and the literature, ten species occurred only in one bay, albeit at times quite abundant within that bay; these were Amathia brasiliensis, Amathia citrina, Aspidelectra melolontha (Landsborough, 1852), Bugulina fulva, Celleporaria brunnea (though it occurred at multiple sites on the west coast where it is native), Celleporaria umbonatoidea, Conopeum cf. reticulum, Membraniporopsis tubigera, Sinoflustra annae, and Smittoidea spinigera (Liu, 1990) (Table 1). All other species occurred in two or more bays. The most frequently documented species across bays, occurring in multiple bays on multiple coasts, were Cryptosula pallasiana (Moll, 1803), Bugula neritina (presumed to be the shallow water Type S of Fehlauer-Ale et al. 2013), and Schizoporella japonica. Across latitudes, the most widespread species along the Pacific coast were Fenestrulina delicia Winston, Hayward & Craig, 2000, and Schizoporella japonica Ortmann, 1890 which both spanned 28° of latitude from San Diego, California to Valdez, Alaska. On the Atlantic coast the most widespread species was Bugula neritina spanning 16° of latitude from Biscayne Bay, Florida to Narragansett Bay, Rhode Island.

Non-native fouling species that were reported in the literature but did not occur on our plates included Smittoidea spinigera (collected on shell hash from bottom grab samples as part of the soft sediment component of this project and to be published elsewhere), Bugulina simplex (Hincks, 1886), Membraniporopsis tubigera (Osburn, 1940), Victorella muelleri (Kraepelin, 1887), Bulbella abscondita Braem, 1951, and Bugula neritina Type N of Fehlauer-Ale et al. (2013) detected by genetics only (Table 2).

In our surveys, the greatest number of non-native bryozoans were detected on the Pacific coast (37), followed by the Atlantic (19) and the Gulf of Mexico (8). On the Atlantic and Gulf of Mexico coasts, most bryozoan introductions were from some region of the Pacific (Fig. 2), with the highest proportion of those to the Gulf of Mexico coming from the Indo-Pacific. In contrast, the greatest number of introductions to the Pacific coast were from the Atlantic. Many of the introductions to all coasts were of unknown origin (Fig. 2).

On the Pacific and Atlantic coasts, non-native bryozoan richness per bay declined significantly with increasing latitude yet differed between these two coasts (Fig. 3). The Gulf coast was not included in this comparison, due to limited sample size and latitudinal range. Also noteworthy are a few outlier bays, or hot spots, on the west coast with exceptionally high numbers of non-native species, including San Diego, Long Beach, San Francisco, California and Coos Bay, Oregon. The latitudinal pattern holds with and without these outliers (Fig. 3).

Non-native species assemblages in bays differed significantly across each coast (PERMANOVA, p < 0.001, Fig. 4). Many non-native bryozoans on all coasts were unique to one coast, with 25 being unique to the Pacific, 6 to the Atlantic and 2 to the Gulf of Mexico (Table 1).

Figure 4. 

Non-metric multidimensional scaling (nMDS) plot of non-native bryozoan species composition from each of the sampled bays in North America (PAC-Pacific, ATL-Atlantic, Gulf of Mexico-GOM). Each point represents the composition of one bay (based on presence-absence data for non-native bryozoan taxa), with the distance between any two points representing the differences between those two bays’ communities as defined by the Bray-Curtis dissimilarity metric (closer = more similar). nMDS stress = 0.08.

Most introductions from our dataset are likely attributable to hull fouling or ballast water (42/48 = 87.5%), with very few being attributable to accidental transport with oysters (6/48 = 12.5%).

The Order Cheilostomata (calcified species) makes up the greater proportion of non-native species (73%), with the uncalcified Order Ctenostomata making up only 27%. Both latitude and ocean are significant predictors of non-native species richness for the Ctenostomata (p < 0.0001 for latitude and p < 0.02 for ocean) and the Cheilostomata (p < 0.0001 for latitude and p < 0.0001 for ocean) with no interaction.

Most non-native bryozoan species were lecithotrophic (81%), the most common larval form in bryozoans. Morphology was almost evenly divided between encrusting (46%) and upright forms (54%) with 4% of species being both upright and encrusting. Both latitude and ocean are strong predictors of richness for encrusting and upright morphologies (p < 0.0001 for latitude and p < 0.001 for ocean for upright morphologies, p <=0.001 for latitude and p <=0.0036 for ocean with no interaction with morphology). There were no non-native Cyclostomata documented, as their difficult taxonomy resulted in many species being designated cryptogenic (Suppl. material 1).

Shipping, in the form of both hull and other external niche fouling, and ballast water, has been the primary mechanism for introducing marine and estuarine bryozoans to North America. Commercial oysters transported from Japan and from the U.S. Atlantic coast to the Pacific coast have led to the introduction of several additional species. While most bryozoans have lecithotrophic larvae with limited natural dispersal capacity, interoceanic, transoceanic, and coastal dispersal by shipping and aquaculture have significantly altered the distribution of a vast number of species. In addition, both transoceanic and coastal rafting on marine debris now provide a significant new transport vector for bryozoans (McCuller and Carlton 2018; Murray et al. 2019; Haram et al. 2020, 2023; Brandler and Carlton 2023).

Several studies have shown that overall species diversity decreases towards the poles (Rohde 1992). Our data on non-native bryozoan diversity also follows this pattern. Polar regions historically are less invaded for many reasons, including the presence of less ship traffic, lower propagule pressure, lower temperatures, and lower rates of anthropogenic disturbance (Sax 2001; Ruiz and Hewitt 2009). However, with ocean temperatures increasing globally, particularly at the poles, this pattern is likely to change as more species are able to survive the warming conditions in northern regions. Additionally, warming waters have also facilitated the opening of ship traffic through the Northwest passage, with new and enlarged port facilities recently funded in the Arctic port of Nome (Thiessen 2023) in anticipation of increased traffic through the Arctic. This is particularly important given that non-native tunicates and bryozoans were transported primarily by shipping related mechanisms (Table 2). Ruiz et al. (2011) found that shipping was the primary transport vector in some of the mobile taxa (crustaceans and molluscs) as well. This suggests that shipping, particularly hull fouling, needs the most focus for preventative management of invasive species across a wide range of taxa, including the bryozoans. Fortunately, analysis suggests that shipment pre-screening or cleaning of vessels can be cost effective (Keller et al. 2006) and decrease the likelihood of introduction (Davidson et al. 2008).

Conclusions relative to sources of non-native species on the Atlantic coast may be confounded by the fact that many North American Atlantic coast species with European names are considered native species. Given that ship traffic across the North Atlantic commenced more than 350 years ago, long before intensive biogeographic studies in North America, it seems likely that some, if not many, of these species may have been introduced since the 17^th century but are overlooked as invasions (Carlton 2003). Interestingly, the only species we documented traveling from west to east (Pacific to Atlantic) was Celleporaria brunnea.

Bryozoan introduction patterns parallel those of tunicates and sessile polychaetes in that more non-native species introductions have occurred to the Pacific coast compared to the Atlantic and Gulf coasts (Simkanin et al. 2016; Bastida-Zavala et al. 2017). Both non-native bryozoan and tunicate diversity decrease with latitude, but sessile polychaetes do not follow this pattern, with more polychaete taxa found in temperate latitudes. The lack of concordance for a latitudinal pattern of non-native sessile polychaetes noted here is consistent with several studies which suggest that polychaetes typically do not follow latitudinal trends, being more tied to local ecological factors (Haanes and Gulliksen 2011; Capa et al. 2021).

As with bryozoans, the majority of introductions to the Atlantic coast among tunicates and sessile polychaetes were from the Pacific, and most introductions to the Gulf of Mexico were from the Indo-Pacific. However, the dominance of Atlantic-sourced invasions to the Pacific coast in bryozoans contrasts with tunicate and sessile polychaete invasions, which are predominately from other Pacific source regions.

There are many more non-native marine bryozoans (48 species) detected in North America than tunicates (27 species) and sessile polychaetes (9 species). The lower number of the latter may be due to increased scrutiny over time of the former two groups. Among the bryozoans, and the tunicates, although to a lesser degree, hot spots of high non-native species diversity are found in the Pacific coast ports of San Francisco, CA, San Diego, CA, Long Beach, CA, and Coos Bay, OR (in order of number of non-native species present, see Fig. 3). All four of these coastal areas have marine labs and are regions that have been intensely studied by researchers with expertise in taxonomy and non-native species. The three “hot spot” California bays are large shipping ports, contributing to the large number of non-native species. Although the number of commercial vessels entering San Diego Bay is significantly lower than those entering San Francisco Bay and Long Beach (National Ballast Information Clearinghouse 2023), it is also home to a US Navy fleet that spends long periods of time in port, increasing the risk for hull fouling introductions. San Francisco Bay, California, with the second highest shipping traffic on the west coast overall, has a higher proportion of non-native species than any other North American embayment of similar size, and a higher proportion of first records for non-native species in all taxonomic groups, accounting for 65% of the species introduced to the Pacific coast (Ruiz et al. 2011). Historically high propagule pressure to the Bay from multiple vectors may help to explain San Francisco Bay’s high numbers of non-native species (Ruiz et al. 2011). Not surprisingly, the port with the highest shipping traffic on the west coast, LA/Long Beach, is also one of the hot spots of marine invertebrate biodiversity, including that of introduced bryozoans, however the number of non-natives is much lower than in San Francisco and San Diego Bay. This may be attributable to the fact that salinities are fairly high and stable in the Long Beach area, ranging from 28–35 PSU at our sample sites, while the range for San Diego (21–37 PSU), San Francisco (0–36 PSU) and Coos Bay (10–32 PSU), show wider variability in salinity, affording a greater range of habitats for potential invasions. Coos Bay also had fewer non-natives than San Francisco and San Diego, but it is a much smaller bay. We suggest that the diversity of fully marine invaders may be lower than the diversity of invaders that can live in both marine and estuarine waters.

The total number of non-natives detected increased with repeated sampling in 8 bays sampled at least twice. A few species disappeared and reappeared in repeated years of sampling at a site (Chang et al. 2018, 2020). Some of the non-natives detected appeared only in one year or were represented by only one specimen, further underscoring the value of repeated sampling to determine whether a species is established. For example, only a single specimen of Bugula neritina was detected in Ketchikan, Alaska in 2016, but yearly citizen science monitoring and a repeat settlement plate survey there in 2021 has not turned up the species again (Jurgens et al. 2018; Ruiz unpublished data).

This work strongly supports the use of settlement plates to detect non-native species in coastal ecosystems. A recent analysis found that of 327 non–native marine and estuarine species in North America, 71% of them were reported from hard substrates and most of those were from artificial structures like docks and marinas (Ruiz et al. 2009). The vast majority of non-natives currently known from North America were detected using this methodology.