A doubled down invasion of the northeast Pacific by the Asian mud shrimp, Upogebia major and its coevolved bopyrid isopod parasite, Orthione griffenis

Dramatic declines of the native northeast Pacific mud shrimp, Upogebia pugettensis over the last three decades have occurred in response to intense infestations by the Asian bopyrid isopod parasite, Orthione griffenis, that was introduced in the 1980s. We report herein the arrival of the Asian mud shrimp, Upogebia major, in San Francisco Bay no later than 2006. Complications of identifying juvenile U. major and inefficiencies of collecting mature and readily identified specimens recovered by conventional sampling devices are likely to have delayed its identification and discovery. U. major is less vulnerable to O. griffenis and is displacing or replacing U. pugettensis in its present 200 km range to the north and south of San Francisco Bay. Upogebia major, as a coevolved alternative host, assures persistence of O. griffenis in this region even where native species extinctions occur and can potentially expand to all habitats that are presently invaded by O. griffenis (Alaska to Baja California Norte). The individual and combined O. griffenis and U. major invasions thus threaten U. pugettensis in particular and all other native Upogebia species occurring north of Mexico. Our review of Upogebia taxonomy for a key to species revealed a previously reported 1912 invasion of San Francisco Bay by Upogebia affinis that was in error; hence, the introduction of U. major is the first confirmed gebiid invasion in the world. Greater resolution of U. major natural history and timing of its invasion is needed to test whether it evaded present vector management efforts. Intervention is warranted to limit the doubled down U. major and O. griffenis invasion and to conserve U. pugettensis and other native Upogebia species from ecological or absolute extinction in the coming decades.


Introduction
San Francisco Bay, California is among the most intensely invaded and modified estuaries in the world (Cohen and Carlton 1998;Nichols et al. 1986). Moreover, nonindigenous species of Northeastern Pacific estuaries pose significant threats to human health, the economy, and the environment and attempts to eradicate these species after they become established are costly and often unsuccessful (Scianni et al. 2019). Vector management remains the most effective first defense from the ecological and economic costs of nonindigenous species (Williams et al. 2013) while nonindigenous species that can escape from vector management efforts and that have the greatest environmental or economic effects are of the greatest concern (Lowe et al. 2000).
Upogebia pugettensis (Dana, 1852) was the only confirmed Upogebia species occurring on the western North American coast north of Morro Bay, California (Williams 1986). Upogebia pugettensis live in permanent mud lined burrows (Thompson 1972) even in rocky areas, and form dense burrow galleries in intertidal estuary mudflats. Upogebia pugettensis are classified as ecosystem engineers (D'Andrea and DeWitt 2009;Dumbauld et al. 2021) due to their intertidal burrows and burrowing activities in estuary mudflats. All Upogebia, including U. pugettensis, expand effective estuary sediment surface areas (Chapman and Carter 2014) and circulate oxygenated water deep into suboxic subsurface sediments (Thompson and Pritchard 1969;Thompson 1972). Upogebia burrowing accelerates sediment remineralization and can dominate estuary biogeochemistry (D'Andrea and Dewitt 2009;Pascal et al. 2019). Upogebia pugettensis also support broad diversities of obligate and facultative commensal species . Alterations of major burrowing shrimp populations are therefore also major alterations of the estuary ecosystems that they dominate (Dumbauld et al. 2021).
All major U. pugettensis populations that have been examined (southern British Columbia to Morro Bay, California) since the early 2000s have declined dramatically due to effective castration and reproductive losses to the introduced Asian bopyrid isopod parasite, Orthione griffenis Markham, 2004(Smith et al. 2008Griffen 2009;Dumbauld et al. 2011Dumbauld et al. , 2021Chapman et al. 2012;Repetto and Griffen 2011;Chapman and Carter 2014;Asson et al. 2017;Dumbauld and Bosley 2018). Chapman et al. (2012) concluded that U. pugettensis were extinct in the San Francisco Bay area after they did not find them in multiple intertidal surveys. Chapman et al. (2012) did not survey the subtidal benthos, where U. pugettensis have seldom been reported. The severe declines of previously abundant intertidal U. pugettensis in San Francisco Bay area and all other estuaries from British Columbia to Morro Bay, nevertheless, places O. griffenis, the species responsible for its extirpation, also, as an ecosystem engineer.
We discovered a new Upogebia species in San Francisco Bay from two mature males recovered from San Pablo Bay (in San Francisco Bay, California) in sediments pumped from a 10 m depth area by US Army Corps dredge, Essayons, on 6/10/2010 (McGowan 2010;McGowan et al. 2012). We tentatively identified these subtidal Upogebia, that were the second most abundant invertebrate in the Essayons spoils and that were clearly not the native U. pugettensis, as the Asian Upogebia major (de Haan, 1841). Risks of overestimating the distribution ranges of species or of incorrectly assigning species with the status of "non-indigenous" occur however, where identifications are not confirmed by thorough taxonomic investigations (Simon et al. 2019). We therefore attempted a more conclusive identification of this newly discovered species, which required female specimens and a review of the world taxonomy of Upogebia. Our review was needed also to test previous conclusions of massive U. pugettensis declines, based almost entirely on surveys of intertidal estuary mudflats, and to reexamine a previously assessed "failed introduction" of the western Atlantic, Upogebia affinis (Say, 1818) into San Francisco Bay.
With the exception of U. affinis, the predominantly intertidal U. pugettensis were the only previously reported Upogebia in San Francisco Bay (Ricketts et al. 1985;Dumbauld et al. 1996;Kuris et al. 2007;Jensen 2014). Upogebia affinis were exceptional because Williams (1986) reported them from the 1912-13 US Fisheries Commission research vessel, "Albatross", survey of San Francisco Bay, they were the only subtidal Upogebia collected in the Albatross survey and also, they were the only species other than U. pugettensis reported between Morro Bay, California and Prince William Sound, Alaska (Williams 1986).  included U. affinis in their list of "earlier inoculations into the San Francisco Bay estuary". Fofonoff et al. (2021) proposed this same U. affinis record was "a failed introduction" (introduced but did not sustain reproductive populations).
Whether the mature, abundant Essayons Upogebia were an overlooked native species or a recently exploding population of cryptic origins and history were, initially, open questions. Upogebia have been relatively rare in the last 100 years of samples from San Francisco Bay (Shaeffer et al. 2007, herein). Nearly all sampling from the subtidal San Francisco Bay benthos in the last 100 years however, has been by conventional devices which seldom penetrate more than 20 cm below the sediment surface (Nichols 1973). Upogebia can burrow to more than two meters below the sediment surface (Thompson 1972;Kinoshita et al. 2002) where they are beyond the reach of conventional sampling devices. The up to 28.6 m reach of the Essayons is deeper than all other benthic sampling devices and is therefore more suited for collecting deep burrowing subtidal Upogebia.
The abundant Upogebia from the Essayons spoils thus potentially contradicted Chapman et al.'s (2012) conclusions of local U. pugettensis extinctions. Depending on how long these new Upogebia had been overlooked, their discovery opened to question whether the introduction of the western Atlantic U. affinis into San Francisco Bay was a "failed" introduction or whether another new invader had escaped through existing vector management efforts. The Upogebia collected by the Essayons, if verified to be of Asian origins, were also significant as a previously unrecognized alternative host for O. griffenis that could assure persistence of O. griffenis even where they extirpate native Upogebia species. We partially address these questions in this report.

Materials and methods
We reviewed nearshore northeastern Pacific and global Upogebia taxonomy to clearly identify and distinguish the Essayons Upogebia from all other Upogebia species of the world. Our review included the natural history, ecology and population biology of northeastern Pacific nearshore Upogebia to assess the additional risks of this new Upogebia to native species if it proved to be introduced. We surveyed Upogebia populations in major California estuaries to assess the range of this new species, to establish a baseline for tracking its present and future distribution and to verify the distributions of intertidal Upogebia species north of Mexico that are likely to occur within its range. Our field surveys included direct sampling, correspondences, personal communications, published and gray literature, and on-line postings and databases (i.e. GBIF 2019; iNaturalist 2019; UMLAW 2019; Fofonoff et al. 2021;Ahyong et al. 2020). We also surveyed invertebrate collections in the Los Angeles County Museum of Natural History, Los Angeles, CA (NHMLAC), the California Academy of Sciences, Invertebrate Zoology, San Francisco, CA (CASIZ), the Smithsonian Institution, Invertebrate Zoology, National Museum of Natural History, Suitland, MD (NMNH) the Moss Landing Benthic Lab., Moss Landing Marine Laboratory, Moss Landing, CA and the California Department of Fish and Wildlife, Stockton, CA (CDFW 2019) reference collections.
We compared the Essayons Upogebia with the other recognized nearshore coastal California species, Upogebia macginitieorum Williams, 1986, with U. pugettensis with Upogebia affinis (as conceived by Williams 1993) and with the 4 San Francisco Bay U. affinis specimens reported in Williams (1986). Williams (1993, p. 5) reasserted the possible East Pacific introduction of U. affinis. Upogebia felderi Williams, 1993, of the Western Gulf of Mexico and U. paraffinis Williams, 1993, of Ceará to São Paulo, Brazil, that Williams separated from U. affinis narrowed the potential sources of the San Francisco Bay population to the North Atlantic. We reexamined the San Francisco Bay U. affinis collection notes and data to verify their specific status within the new U. affinis complex and their north Atlantic origins.
We included a photo illustrated key and notes in our correspondences and presentations ) that is included herein to facilitate identifications and public outreach. Photos in the amended key are from a Canon 5D camera focused through an MP-E 65 mm macro lens or through an EF 24-105 mm macro lens attached to a stereo-microscope at 1-20X magnifications. Depth of field was increased by Helicon Focus assemblies of composited, stacked images. We also summarize confirmed and critical Upogebia species records in Figure 1, Supplementary material Table S1, confirmed plus reported U. pugettensis, U. macginieorum, O. griffenis and the native bopyrid, Phyllodurus abdominalis Stimpson, 1857 in Table S1. The photographs of U. affinis from Oyster Bay Harbor, Oyster Bay, New York, in the key ( Figure 2) are in place of our low-quality photographs of the San Francisco Bay U. affinis specimens in the NMNH collections ( Figure S1).
We sampled for U. major, U. macginitieorum and U. pugettensis among California estuaries by yabby pump (Manning 1975), in low density populations (~< 10 m -2 ) and by hand held cores or shoveled excavations among higher density (~> 10 m -2 ) populations. Appendages of all field collected specimens were preserved on the day of recovery in 70% ETOH for genetic analyses. Voucher specimens from our surveys are deposited in the NHMLAC, CASIZ and the Museum of Natural History, University of Colorado Boulder, Boulder, CO (UCM) museum collections and our notes and reported records of field collected shrimp are summarized in Table S1 and Appendix 1. Field collected shrimp were rinsed of adhering sediments and examined "fresh" or frozen (to fix and prevent escapes of mobile parasites). The shrimp were then thawed and inspected directly for bopyrid isopod parasites and other external commensals, then washed to dislodge parasites under the carapace. The wash water was collected into a 100 mm petri dish where it was examined under 6-10X magnification for small commensals and parasites. We distinguished Upogebia sexes from the first pleopods (present on mature females and absent on mature males), and from the female gonopore (on the distal, medial coxa of female pereopod 3) and the male gonopore (on the distal, medial corner of the coxa of male pereopod 5). Upogebia bearing both male and female gonopores were classified as hermaphrodites. We estimated size from carapace length (CL) that we measured to the nearest 0.5 mm from the distal tip of the rostrum to the posterior carapace edge. Processed specimens were refrozen or transferred into 70-95% ETOH for long-term preservation.
We used photographs ( Figures S2 and S3) to verify juvenile specimens and specimens that were not collected. Our detailed review of San Francisco Bay U. pugettensis from 1878 to 2020 is included in Appendix 2.

Results
The similar morphologies, sizes and colorations of U. major and U. pugettensis can confuse non-specialists, and initially, ourselves. Gyo Itani, Seto Marine Biological Laboratory, Kyoto University, Shirahama, Wakayama, Japan  Table S1). confirmed our initial identification of the Essayons specimens as U. major from the chela, rostrum, merus and pereopod 2 and 3 merus morphologies described in Sakai's (2006) key to species, Kang et al. (2007) and Hong and Lee (2014). The three nearshore shallow water eastern Pacific Upogebia species that can be confirmed north of the California-Mexico border do not include U. affinis.

(Figure 2)
Systematics -The exoskeleton of the abdomen, telson and left uropod are all that remain of the Say's type U. affinis (Williams 1993). Williams (1993) therefore re-described and illustrated U. affinis using a specimen "of size comparable to Say's type when originally described, from South Carolina near the Georgia state line". Williams (1993) separated U. felderi, Williams, 1993, from the Gulf of Mexico, and U. paraffinis, Williams, 1993, from Brazil from the originally conceived U. affinis by: 1) absence of a strong anteriorly projecting spine near to the dorsal margin of the rostrum in U. affinis that is present and, often prominent in U. paraffinis; 2) stronger spines on the chelipeds of U. felderi and U. paraffinis than in U. affinis; 3) strongly developed ventral spines on coxa 1 through 5 of U. affinis that are reduced in U. felderi and U. paraffinis and; 4) granules or short spines on the distal margin of U. affinis and U. felderi uropods that are absent on U. paraffinis.
Description -Rostrum bearing large post-ocular spine (Figure 2A, B) and 0-8 ventral spines ( Figure 2B, bracket). Pereopod 2 merus with proximoventral spine, 1 subdistal dorsal spine (Figures 2D and S1B) and prominent ventral coxa spines (Figure 2A, C). Prominent spines on ventral coxae of pereopods 1-5. Male and female dactyl and thumb similar ( Figure  2E, F). Fine teeth on proximal dorsal thumb of cheliped opposed by a narrow ridge of similarly fine teeth on ventral proximal dactyl ( Figure 2E, F). Ventral medial tooth of the dactyl extending to medial side of the closed thumb.
Remarks -As expected, the four San Francisco Bay U. affinis of William's (1986) match William's (1993) conception of U. affinis and are distinguished from all northeastern Pacific Upogebia species by their ventral distal rostral spines, which are absent on all North Pacific Upogebia. The proximal medial ventral spine on the pereopod 2 merus of U. affinis are in common with U. major and the small teeth on the inner face of the U. affinis cheliped thumb (Figures 2E and S1A) could be misinterpreted by nonspecialists as bare, and thus, confused with the unadorned inner face of the U. major cheliped thumb ( Figure 3C, D). However, our examination of William's San Francisco Bay U. affinis specimens, collection notes and original collection data revealed that the record itself is in error. Williams (1986) reported: "The specimens" (of U. affinis) "may represent an introduction" and further noted: "Carlton (1979)" . . . (the earliest review of introduced marine and estuarine invertebrates of the North American Pacific coast) . . ."did not list this species from San Francisco Bay and to my knowledge no other specimens have been taken from there." These specimens would thus have been remarkable as the only record of the species in the Pacific, as a possible new introduced species on the coast, and because they were from the subtidal benthos.
Other than a male U. pugettensis collected in June 1942 from offshore of Winchester Bay, Oregon (Williams 1986) and one specimen collected in July 1904 from Prince of Wales Island, Alaska (Appendix 1), we are unaware of subtidal northeast Pacific Upogebia records from outside of San Francisco Bay. Other than the 1912 U. affinis, subtidal Upogebia were not reported from San Francisco Bay between 1878 and the 1980s (Schemel et al. , 1990. Schmitt (1921) recognized only U. pugettensis in the East Pacific and emphasized (p. 10) that U. pugettensis was one of only 5 decapod species that were obtained entirely at shore (intertidal) stations from the Albatross survey. Schmitt (1921, p. 115) additionally noted that the "three", "shore" specimens of Upogebia "were taken in the course of the Albatross survey, one at Sausalito, February 8, and two at Tiburon, April 29, 1913", in a "150 foot seine" (p. 353). The validity of Williams' (1986) remarkable "four" sui generis U. affinis from Albatross survey trawl "5824A" require that they were misplaced or overlooked for 64 years despite extensive, meticulous, comprehensive original survey notes (USBF 1921;Schmitt 1921).
The 12/18/1912, Albatross "D" for dredging station "D. 5824A" sample was part of a modified 5 foot wide Tanner beam trawl "towed approximately 2 km between orange peel bucket dredge stations 5824A, 7 fathoms and 5824B, 12.5 fathoms" (USBF 1921, p. 64). The trawl center was approximately (37.9354; −122.4507) (Figure 1, purple line) (also from Plate 1, and p. 338 of Schmitt [1921]). Sediments in these samples ranged from "Soft, gray mud, slightly gritty, not very tenacious, at beginning and shelly, somewhat muddy sand, at end of haul." Schmitt's (1921, pp. 338 (Schmitt 1921). The as yet undetermined origins of Williams' (1986) U. affinis specimens do not include San Francisco Bay. The absence of "Upogebia" (that would have been the U. affinis specimens) from the original collection notes (USBF 1921;Schmitt 1921), the recovery of 4 mature U. affinis from only 1 of 285 benthic samples in the Albatross survey in which crustaceans were recovered (Schmitt 1921) and the lack of additional U. affinis records from San Francisco Bay or outside of the west Atlantic in the following 7 decades preceding Williams' (1986) report, place the source location of these U. affinis specimens in extreme, highly significant, doubt.
Distribution -Upogebia affinis are confirmed only from the east coast of North America, the Western Atlantic, from Massachusetts to southern Texas. Description -Rostrum smoothly rounded ventrally and lacking ventral teeth. One or two ocular spines ( Figure 3A, red arrows). Thumb of male and female chela propodus length 1/3 of dactyl length and lacking teeth or tubercles on inner and outer distal surfaces ( Figure 3C, D). Cheliped dactyls strongly dimorphic between sexes. Male cheliped dactyl medial surfaces bearing three prominent oblique ridges and a small lateral tooth ( Figure 3C). Female dactyl with two longitudinal rows of tubercles ( Figure 3D). Pereopod 2 merus with one proximal ventral spine and two large proximal medioventral spines ( Figure 3E). Weak spines on ventral coxae of pereopods 1-3 ( Figure 3B). Upogebia major are similar in general morphology and color to U. pugettensis but mature U. major are readily distinguished in the field by the smooth, unadorned the cheliped thumb inner face ( Figure 3C, D). Sakai's (2006) key to species, illustrations in Kang et al. (2007) and photographs in Hong and Lee (2014).
North American Distribution -Bodega Harbor, Drakes Estero, Tomales Bay, Bolinas Lagoon, San Francisco Bay, and Elkhorn Slough, California. Not found in 2/10/2020 to 2/16/2020 searches of San Diego Bay, Mission Bay, Newport Bay, Los Angeles Harbor or Morrow Bay (Figure 1, Appendix 1, Table S1). Not found among extensive annual collections of U. pugettensis from Oregon or Washington over the last decade (Smith et al. 2008;Dumbauld et al. 2011;Chapman et al. 2012;Chapman and Carter 2014;Asson et al. 2017;Breitenstein et al. 2017;Li et al. 2020). The 200 km (Elkhorn Slough to Bodega Harbor) range of U. major (Figure 1) includes all major estuaries adjacent to San Francisco Bay, where it is most likely to have first established.
Remarks -Agonistic behaviors of in-hand specimens or among multiple U. major held in the same containers are minor. The prevalence of solitary or multiple tube occupancy is unclear. The roughly 2:1 U. major F:M sex ratio (herein) is consistent with hermaphroditism.
North American U. major are a possible "Category 4" introduced species "misidentified initially as a known native species" (Carlton 2009). Prereproductive U. major, that are most accessible to conventional quantitative sampling devices, are readily confused with U. pugettensis. Mature U. major, which are more readily identified, can burrow to 2 m depths (Kinoshita 2002) that are beyond the reach of conventional sampling devices. The dredged Essayons sediments have seldom been examined by biologists. Upogebia major could thus readily have escaped notice, even in dredge spoils, for multiple years previous to McGowan's discovery and at least one Upogebia species was abundant in San Francisco Bay in the early to mid-1980s.
Associated species -Three of the commensal clams, Neaeromya rugifera (Carpenter, 1864) and two solitary juvenile O. griffenis (0.9 and 1.4 mm in length), were recovered from 18 U. major collected between 5/5/2018 and 5/6/2018 from Bodega Harbor in an approximately 50 m diameter area of the approximately 1 km 2 overall intertidal survey area. Both U. pugettensis recovered in that survey were from the same 50 m diameter area as the Description -Rostrum smoothly rounded ventrally, lacking ventral teeth. Postocular spine absent ( Figure 4A). Pereopod 2 merus lacking proximal-ventral spine ( Figure 4B). Ventral spines on coxae absent. Male chelipeds ( Figure 4C) more robust than female chelipeds ( Figure 4D), similar overall, with minor row of tubercles extending the length of the medial dactyl face. Powerful tooth (red points) at the base of the dactyl. Dactyl closing over a broad tooth on the thumb that is separated proximally from the main propodus by deep sinus (Figure 4B, F). Dorsal anterior propodus without teeth or spines ( Figure 4A).
Remarks -The vertical distribution of U. macginitieorum extends from low intertidal mudflats to high intertidal Salicornia marshes (Campos et al. 2009; J. Chapman personal observations), cord grass stands adjacent to marsh channels (R. Appy personal observations) and clay banks (Campos et al. 2009). Orthione griffenis recovered from San Diego Bay and Los Angeles Harbor populations. U. macginitieorum not found in 2/15/2020 searches of lower and upper Newport Bay, California. Upogebia macginitieorum female to male ratio (herein) roughly 2:1, consistent with hermaphroditism. Agonistic behaviors minor and burrow cohabitation occurs.

(Figure 5)
Material examined and records -Appendix 1. Description -Small ocular spine ( Figure 5A). Pereopod 2 merus without spines ( Figure 5B). Chela more robust in males ( Figure 5C) than females ( Figure 5D) but similar overall, with strong tooth on dorsal medial thumb ( Figure 5C, red point) and smaller tooth ( Figure 5D, red point) on medial ventral dactyl that does not meet with the closed dactyl ( Figure 5C, D). Ventral coxae of pereopods 2 and 3 lacking spines ( Figure 5E). Williams (1986). Distribution -Previously, Prince William Sound, Alaska to Morro Bay, California (Williams 1986). All known populations between Ketchikan, Alaska and Elkhorn Slough, California are of uncertain status, collapsed or extinct.

Systematics -Summarized in
Remarks -Review of U. pugettensis in San Francisco Bay from the 19 th century to 2020 and our 2020 field survey (Appendix 2) resulted in 6 major findings: 1) Orthione griffenis infestations (and probable effective castration) among U. pugettensis co-occurring with U. major (Bodega Harbor and Tomales Bay) (Appendices 1 and 2) were 9.8 times more frequent than among U. major. The 1:1 U. pugettensis sex ratio (Asson et al. 2017, herein) is consistent with gonochorism, which may require greater densities for reproduction and establishing new populations than are required for the possibly hermaphroditic U. major. Extreme agonistic behaviors among mature U. pugettensis require separation to limit dismembering (personal observations). The greater aggression of U. pugettensis and their maintenance of solitary burrows are consistent with high territory value and territoriality. All known intertidal U. major populations in all estuaries of the San Francisco Bay area occur where U. pugettensis previously occurred.

Shallow water Upogebia from San Diego Bay, California to Alaska
Exclusion of U. affinis from popular guides to East Pacific decapods (i.e., Kuris et al. 2007;Jensen 2014) was logical but, without explanation of the omission, left U. macginitieorum and U. pugettensis as the default Upogebia species north of Mexico. We include U. affinis in our key to prevent its confusion with U. major by non-specialists, including us initially. Upogebia major is distinguished from both confirmed nearshore East Pacific Upogebia species occurring north of Mexico (U. pugettensis and U. macginitieorum) by the combination of smooth, toothless inner face of the cheliped thumb and the proximal medial ventral pereopod 2 merus spine.

Discussion
The combined O. griffenis and U. major invasion threat is greater than their independent sums and places them among the most consequential marine invasions that have occurred in the East Pacific. Orthione griffenis effectively place its coevolved host, U. major, at competitive advantage over every native Upogebia species where they both invade. Upogebia pugettensis within the U. major range occur at lower densities than U. major and carry approximately 10 fold or greater O. griffenis infestation intensities. As an alternative host, U. major assures persistence of O. griffenis even where native Upogebia extinctions occur. Upogebia major, occurring in subtidal and intertidal habits (Figure 1, Appendix 2) overlap the almost exclusively intertidal U. pugettensis and thus are adapted to all known U. pugettensis habitats within its present range. Orthione griffenis and U. major reveal, once again, the long-term values of controlling invasion vectors (vector management, Williams et al. 2013;Ruiz et al. 2015) and of limiting ongoing invasion impacts (Pyšek et al. 2020).
Vector management can most effectively prevent introductions "by managing the many vectors that deliver them, rather than by focusing efforts on the management of individual species or even individual vectors" (Williams et al. 2013). Vector management requires species by species knowledge to identify likely vectors and information on prevention effectiveness. The second defense requires information on invasion risks and possible responses based on estimated present and future ranges, impacts and mechanisms controlling invasions underway.

Vector effectiveness assessed from invasion timing Upogebia major and O. griffenis could have established in San Francisco
Bay together from unsanctioned live seafood releases (Cohen and Carlton 1997;Chapman et al. 2003) or independently as larvae and pre-settlement decapodids, respectively, that arrived in ballast water traffic.  were unaware of U. major in North America when they reported the O. griffenis invasion. They inferred that O. griffenis directly caused the collapse of San Francisco Bay U. pugettensis that were abundant up to the late 1980s (Schemel et al. , 1990. Information on whether U. major can re-burrow and the coincidence of O. griffenis and U. major invasion dates will be critical for determining the timing of these invasions and their likely seafood or ballast water introduction vectors. Upogebia major infested with O. griffenis could have been introduced to San Francisco Bay together as live seafood or bait. The Chinese mitten crab, Eriocheir sinensis Edwards, 1853, and the green crab, Carcinus maenas (Linnaeus, 1758), were introduced initially into San Francisco Bay as unsanctioned releases of, respectively, seafood  and bait (Cohen and Carlton 1997). Upogebia major are a popular Asian seafood (Hooper 2006;Inoue et al. 2002;Inoue et al. 2008). Upogebia pugettensis were a popular bait species (Chapman and Carter 2014) and live seafood species are commonly purchased in North America for consumption but released later as bait (Chapman and Miller 1999;Chapman et al. 2003). The probabilities of both seafood or bait vectors depend on whether U. major are competent to reburrow. Upogebia cannot survive outside of their burrows where they are vulnerable to predation by fish ) and birds ( Figure S2C; Stenzel et al. 1976). Released mature Upogebia can thus survive only by quickly re-burrowing. Burrowing competence of U. major however, is unknown. Reproductive size U. pugettensis cannot reburrow (Thompson 1972;Jackson et al. 2011; personal observations) while other Upogebia species can (Pascal et al. 2019).
A likely alternative transoceanic vector of U. major and O. griffenis would be as larvae entrained in transoceanic ballast water traffic. U. major and O. griffenis larvae would not likely be transported in the same seasons or years. Upogebia major produce larvae, in late winter that settle by late spring (Konishi 1989). Orthione griffenis produce larvae through summer that settle in fall (J. Chapman personal observation). The first in-hand O. griffenis place them in North America by the early to mid 1980s ). The first in-hand U. major was a reproductive size, 30.5 mm CL, (2-5 year old) male from CDFW station 103 in South San Francisco Bay, collected in December 2009 ( Figure 1, Table S1, Appendix 1) (that was initially misidentified as U. pugettensis). The next earliest in-hand U. major were the two 6/10/2010 (23.4 mm CL and 33.5 mm CL) no less than 2 year old reproductive size Essayons males. Depending on their ages, these early specimens place U. major in San Francisco Bay between 2004 and 2007, with widespread populations no later than 2008. A conservative "earliest in-hand estimate" of U. major presence from these specimens, based on their minimum likely 2+ year ages is at least by 2006. Upogebia major would thus have been overlooked for around 20 years if O. griffenis was a seafood or bait introduction among infested U. major. A 20 year discovery gap for these two invasion thus, would be remarkable if they came with seafood or bait but not if they came as larvae in ballast water. A later introduction of U. major via ballast water traffic however, would have occurred after ballast water restrictions were in place.
Ballast water treatment and seafood vector management efforts were minor before the 1990s. Greater ballast water or seafood management efforts may thus be warranted only if U. major established in the East Pacific in the early to mid-1980s with O. griffenis before management efforts were in place . The State of California, effective Jan 2000, mandated the exchange of ballast water from ships traveling outside the US Exclusive Economic Zone and entering California ports to stem future biological invasions (Scianni et al. 2019). The California Marine Invasive Species Act of 2003 extended the Ballast Water Management Act of 1999 to prevent the discharge of nonindigenous species into the waters of the state (CDFW 2008(CDFW , 2011Scianni et al. 2019). San Francisco Bay nevertheless, received nearly 3 million m 3 of "treated" (exchanged in open ocean) and untreated ballast water from foreign countries between January 2000 and June 2002 (Choi et al. 2005). Surveys of commensal clams could narrow the 20-year knowledge gaps between when the U. major invasion began (in the 1980s or after 2004) and whether U. major are likely to have arrived when vector management efforts were in place.
Marine mollusk shell deposits can indicate mollusk population dynamics over time (i.e., Schopf 1978). The native clam, Cryptomya californica (Conrad, 1837) is commensal with U. major and U. pugettensis. Deposits of C. californica in the overlapping U. major and U. pugettensis distributions (Figure 1) may indicate the abundances of these two shrimps through time, even where they were not collected. Analyses of C. californica shell remains could thus address the decades wide information gaps between when O. griffenis invaded and decimated U. pugettensis populations and when U. major established and became abundant. Cryptomya californica are abundant in the burrows of all mature native and introduced Upogebia species in California and north and are readily collected in conventional sampling devices. The distributions and size structures of live and deposited C. californica valves collected in past or new surveys can reveal the sizes, ages and distributions of their extinct or recently arriving shrimp hosts.
Cryptomya californica, for example, were absent from stations D41A-C "San Pablo Bay, near the mouth of Petaluma River-Center" (38.053, −122.4138) between 1996 and 2006 and present in all following years (IEP 2019) ( Figure S4). Upogebia major were not collected from San Pablo Bay until June 2010. A year or more would have been required after Upogebia arrived for C. californica to settle and grow to large enough sizes to be collected and identified in the IEP survey. Upogebia therefore established at the IEP D41A-C station a year or more before C. californica could appear. The appearance of C. californica in 2007 is roughly consistent with the 2006 estimate from our 2009 in-hand U. major from South San Francisco Bay (Figure 1). The presence of C. californica also, again, place U. major as widely established and abundant in San Francisco Bay no later than 2007 or 2008 but no earlier than 2006.
The absence of C. californica at stations D41A-C between 1996 and 2006 ( Figure S4) also place the decline or extinction of U. pugettensis in San Francisco Bay previous to 1996, when O. griffenis are likely to have been present for at least a decade, and before the first "in-hand evidence" of U. major in the bay at around 2006. Aging live and deposited C. californica shells at this and other San Francisco Bay area stations would provide another method for the invasion sequence analyses.

Assessing risk
Estimates of present and future impacts and knowledge of how these impacts occur or could be mitigated are needed for a second defense against established and expanding invasions. The impacts of this double invasion appear to be severe. All of our records of U. major in Drakes Estero, Bolinas Lagoon, San Francisco Bay and Elkhorn Slough mudflats are from where U. pugettensis were abundant but also no longer occur. Orthione griffenis infest a broad diversity of, and possibly all, Upogebia and Austinogebia species within its Asian range (Itani 2004;Chapman et al. 2012;Hong et al. 2015). Orthione griffenis have infested all U. pugettensis populations and most U. macginitieorum populations examined over its 3,000 km Alaska to Punta Banda, Baja California Sur, Mexico range (Campos et al. 2009;Chapman et al. 2012 and herein). The less vulnerable U. major are a refuge for O. griffenis, that assures its persistence even where extinctions of native Upogebia species occur. Conservation, a second level defense, will benefit from information on whether or when Upogebia populations within the present O. griffenis range will not be invaded by U. major.
The present 200 km range north and south of San Francisco Bay place the initial North American origins of U. major within San Francisco Bay. San Francisco Bay is surrounded by the greatest human populations in the region and is the only international shipping port among the presently invaded estuaries (Figure 1). Estuaries outside of San Francisco Bay receive little or no ballast water traffic and are likely to have been invaded by pelagic U. major larvae that dispersed from San Francisco Bay via coastal currents. The geography and order of U. major collection records indicate they are spreading away from San Francisco Bay (Figure 1). We did not find Upogebia in searches of the extreme north end of Tomales Bay or in Bolinas Lagoon in 2008   The present 3,000 km East Pacific (Ensenada, Baja California Norte, Mexico)  to Ketchikan, Alaska (herein) range of O. griffenis is likely to have established via larval dispersal in coastal currents. The 2,000 km Morro Bay, California to Haida Gwaii, British Columbia (BC) range of the invading green crab, Carcinus maenas, is also likely to have established by larval dispersal in coastal currents. These two invasions may thus be harbingers of an expanding U. major invasion by larval dispersal and consequent collapses of U. pugettensis based ecosystems as the invasion progresses.
Climate diversity within the Asian range of U. major and O. griffenis (southern Japan and central China to Vladivostok, Russia) (Sakai 2006;Sato et al. 2013;Hong et al. 2015;Golubinskaya et al. 2016) exceeds the Morro Bay to Prince William Sound climate range (McGowan et al. 1998;Chapman 2000) of U. pugettensis. Orthione griffenis have been in southern California for more than a decade ) and were established on Calvert Island, BC by 2017 (Whalen et al. 2020) and Ketchikan, Alaska by 2018 (herein). The C. maenas reproductive season and life history (Yamada et al. 2017) are similar to those of Upogebia species (Dumbauld et al. 1996). Carcinus maenas was initially introduced to San Francisco Bay in the 1980s , reached Morro Bay before 2010 (Hall 2010), the Salish Sea by 2016 (Yamada et al. 2017) and the Haida Gwaii Islands by 2020 (Strait of Georgia Data Center 2020). Without intervention, the range of U. major could equal those of O. griffenis and C. maenas within the next 30 years.

Ecosystem shifts
Vectors were necessary but were unlikely to have been the only mechanisms that determined the U. major and O. griffenis invasions. Both species arrived decades or a century after their likely and unregulated ballast traffic and/or the seafood trade dispersal vectors were in place. Both invasions are consistent with increasing vulnerabilities of northeastern Pacific nearshore ecosystems since the 19 th century. Orthione griffenis are most likely to have established initially in an international port, such as San Francisco Bay or in the Salish Sea, years before its 1988 discovery in Willapa Bay, Washington. Willapa Bay receives little or no international ballast water traffic . Orthione griffenis prevalence increased exponentially in Willapa Bay after 1997 ), a decade long invasion lag (Crooks 2005) after they were present. The 1998 and later population explosion of O. griffenis in Willapa Bay coincided with northeast Pacific ocean warming events after 1998 (Cloern et al. 2007;Cloern and Jassby 2012). The European green crab, Carcinus maenas, similarly established in Oregon and Washington estuaries by 1997 (Behrens- Yamada et al. 2005) and thrived in inlets of the west coast of British Columbia since 1998 (Yamada et al. 2017). The major but irregular expansions of C. maenas coincided with ocean warming events , 2005-2006, 2014) (Behrens-Yamada et al. 2005Yamada et al. 2017;Schooler et al. 2020) and their expansions are consistent with 20 th and 21 st century disturbances in San Francisco Bay and northeastern Pacific estuaries.
Subtidal U. pugettensis were vanishingly rare or absent in San Francisco Bay at the time of the 1912-13 Albatross survey (Schmitt 1921, Appendix 2), abundant or common before 1878 (Lockington 1878) and abundant or common again by the 1970s (herein). The low abundance or absence of subtidal U. pugettensis in 1912 coincided with massive alterations of the San Francisco Bay sediments due to hydraulic gold mining of the previous century in the Sierra Nevada that was outlawed in California in 1884 because the derived sediments were blocking navigation channels of the Sacramento-San Joaquin Rivers (Hedgepeth 1979). The greatest quantities of the eroding mine tailings migrating through San Francisco Bay in the early 1900s coincided with the Albatross survey (Hedgepeth 1979). Sediment disturbances of the early 1900s were thus likely to have affected sediment dependent benthic species.
Inverse, sediment dependent interactions of burrowing shrimp and commercially valuable bivalves are well documented (Dumbauld et al. 1996;Pillay et al. 2007;Takeuchi et al. 2013Takeuchi et al. , 2015Hong et al. 2015). The 1912-13 Albatross survey was initiated in response to declining intertidal bivalves (oysters) (Sumner et al. 1914;Nichols 1973) that are widely attributed to manifold human mediated alterations of the San Francisco Bay estuary (Nichols 1973;Cloern and Jassby 2012). Although direct connections between sediment disturbances and declines of oysters and subtidal burrowing shrimp in San Francisco Bay in the early 20 th century were not resolved, later 20 th century sediment disturbances are nevertheless associated with burrowing shrimp abundances and oyster declines.
Pesticide applications to Oregon and Washington estuary mudflats began in the 1950s to manage increased populations of Neotrypaea californiensis and U. pugettensis that were incidentally burying beds of the introduced oyster, Crassostrea gigas (Thunberg, 1795) (Dumbauld et al. 1996. The affected oyster beds had been successfully cultured for the previous 100 years (Dumbauld et al. 1996 on tide flats where oysters had not previously occurred. Those introduced oysters are likely to have altered sediment surface waterflow underlying sediment geochemistry and sediment microflora. Thus, presently unexplained, U. major and O. griffenis invasions followed anthropogenically influenced alterations of animal/sediment relationships. Resolution of the specific mechanisms for these interactions would be useful for oyster aquaculture production and for burrowing shrimp management and conservation. Declines of bivalve biomass and record increases of crangonid shrimp, juvenile Dungeness crab and English sole abundances in San Francisco Bay coincided with abrupt inter-annual and decadal changes in phytoplankton abundances and coastal ocean warming after 1998 (Cloern et al. 2007;Cloern and Jassby 2012) and with the O. griffenis explosion in Willapa Bay . In contrast to the declining native marine decapod species, the introduced catadromous mitten crab, Eriocheir sinensis, (Cohen and Carlton 1989) reached peak abundances in San Francisco Bay in 1998 and declined to near or absolute extinction by 2010 as native decapod populations recovered (Hieb 2011). Except possibly for U. pugettensis (infested by O. griffenis since the late 1980s), subtidal decapod species that have ocean or estuarine dependent larvae, increased in San Francisco Bay after 1998 while free living bivalves declined (i.e. Figure S4). Over the same period, E. sinensis, the only freshwater decapod in San Francisco Bay with estuary dependent larvae, declined. Widespread anthropogenically driven climate and ecosystem shifts that altered U. pugettensis population dynamics and the vulnerability of East Pacific estuaries to invasions have thus been ongoing since at least since the early 20 th century.
The attraction of Upogebia juveniles to adult populations Candisani et al. 2001) and the sharply defined borders of U. pugettensis beds in otherwise seemingly homogenous mudflats (J. Chapman personal observations) indicate that Upogebia are closely dependent on particular sediment compositions and associated microbial communities (Li et al. 2020). Although occasionally found within rocky beaches, all Austinogebia and Upogebia species live in permanent mud lined burrows (Kinoshita 2002;Sakai 2006) constructed of fine sediments cemented together with polysaccharide compounds produced in a hind gut gland, which is the largest organ in the Upogebia body (Thompson 1972). The burrow lining appears to be a critical food source for Upogebia Pascal et al. 2019;Li et al. 2020).
The restriction of U. pugettensis within the present range of U. major to mudflats where they are surrounded by U. major (Bodega Harbor, Tomales Bay and San Francisco Bay) is consistent with overlapping sediment preferences and with high potentials for interspecific competition. Animalsediment relationships in general and the burrowing shrimp-sediment interactions in particular, are thus increasingly in need of resolution for managing and conserving Northeastern Pacific estuary ecosystems.
Changes that exacerbate O. griffenis interactions with U. pugettensis benefit U. major where they co-occur. Additions of species are themselves alterations of the ecosystems they invade. The magnitudes of ecosystem change with species additions can be difficult to measure but default assumptions of the absence of change are false. The effects of U. major in northeast Pacific ecosystems are unlikely to be the same as those created by native Upogebia species that they displace or replace. The doubled down U. major and O. griffenis invasion thus contributes to expanding alterations of northeast Pacific estuary ecosystems (i.e., McGowan et al. 1998;Cloern and Jassby 2012) by presently unmeasured magnitudes and they may facilitate new invasions (i.e., Eriksson and Hillebrand 2019;Blowes et al. 2019;Pyšek et al. 2020). Upogebia major and O. griffenis are also, respectively, the first confirmed introduced gebiid crustacean and first confirmed bopyrid isopod in the world, but are unlikely to be the only burrowing shrimp or bopyrid isopod introductions that have occurred.

Conservation
Intervention is possible and warranted to conserve U. pugettensis and its ecosystems from ecological and absolute extinction. The genetic diversity among the remaining U. pugettensis populations over their North American range could reveal their connectivity, their evolutionarily significant units (ESUs), their populations at greatest risk, and potentially differential resistance to O. griffenis. Genetic information on North American U. major and O. griffenis populations could also indicate their initial North American population sizes and whether they are derived from single or multiple Asian populations. Environmental and bulk DNA samples of whole-communities, coupled with suitably designed primers and DNA metabarcode sequencing (Thomsen and Willerslev 2015;Whalen et al. 2020) could serve to detect remaining or newly invading adult and larval Upogebia species and O. griffenis among estuaries and fjords. These data could be derived from new surveys and from existing DNA analyses and samples produced in ongoing surveys such as the Marine Invasive Species Program and the California Marine Habitats and Estuaries Program (Scianni et al. 2019).