Evolutionary responses of marine organisms to urbanized seascapes

Abstract Many of the world's major cities are located in coastal zones, resulting in urban and industrial impacts on adjacent marine ecosystems. These pressures, which include pollutants, sewage, runoff and debris, temperature increases, hardened shorelines/structures, and light and acoustic pollution, have resulted in new evolutionary landscapes for coastal marine organisms. Marine environmental changes influenced by urbanization may create new selective regimes or may influence neutral evolution via impacts on gene flow or partitioning of genetic diversity across seascapes. While some urban selective pressures, such as hardened surfaces, are similar to those experienced by terrestrial species, others, such as oxidative stress, are specific to aquatic environments. Moreover, spatial and temporal scales of evolutionary responses may differ in the ocean due to the spatial extent of selective pressures and greater capacity for dispersal/gene flow. Here, we present a conceptual framework and synthesis of current research on evolutionary responses of marine organisms to urban pressures. We review urban impacts on genetic diversity and gene flow and examine evidence that marine species are adapting, or are predicted to adapt, to urbanization over rapid evolutionary time frames. Our findings indicate that in the majority of studies, urban stressors are correlated with reduced genetic diversity. Genetic structure is often increased in urbanized settings, but artificial structures can also act as stepping stones for some hard‐surface specialists, promoting range expansion. Most evidence for rapid adaptation to urban stressors comes from studies of heritable tolerance to pollutants in a relatively small number of species; however, the majority of marine ecotoxicology studies do not test directly for heritability. Finally, we highlight current gaps in our understanding of evolutionary processes in marine urban environments and present a framework for future research to address these gaps.


| INTRODUC TI ON
Much of the world's population is concentrated along coasts. More than 40% of the world's population lives within a narrow coastal band, and urban density near the sea is predicted to continue to rise in future decades (McGranahan, Balk, & Anderson, 2007;Small & Nicholls, 2003). Three-quarters of large cities and nearly all megacities (those with >20 million residents) are located on coastlines (UN Habitat, 2009) with extreme consequences for bordering marine habitats. The reach of urban influences, including artificial structures, runoff, industrial dumping, and sewage, extends beyond the terrestrial cityscape to impact coastal organisms and ecosystems. Habitat disturbance or elimination is inherent in waterfront construction including land reclamation, jetties, sea walls, and other forms of shoreline hardening. For example, it is estimated that between one-third to one-half of all vegetated intertidal habitats (e.g., estuaries and seagrass meadows) have been lost globally (Barbier et al., 2011;Duarte, Losada, Hendriks, Mazarrasa, & Marbà, 2013;Perkins, Ng, Dudgeon, Bonebrake, & Leung, 2015). Industries that depend on the marine environment (fishing, marine extraction, shipping) also tend to be concentrated near urban centers. These changes and inputs affect biological communities, with important consequences for processes that drive both marine ecosystem function and microevolutionary processes. However, most studies on urban evolution have focused on terrestrial systems (Alberti et al., 2017;Donihue & Lambert, 2015;Johnson & Munshi-South, 2017;McDonnell & Hahs, 2015;Miranda, Schielzeth, Sonntag, & Partecke, 2013). In contrast, there are relatively few studies that explicitly address the microevolutionary consequences of urbanization in marine ecosystems that fringe cities.
Urbanization results in a highly altered set of selective pressures and environmental variables that may affect demography for marine taxa and vary substantially across the type of impact and the organism. In this review, we consider observed and potential evolutionary changes that are associated with urban environments. While many of the most well-studied examples come from ecotoxicology and the strong selective pressures exacted by concentrated pollutants such as heavy metals, urban impacts such as habitat alteration can affect other aspects of microevolution, such as gene flow and genetic drift.
As in terrestrial systems, urban impacts on marine species are extensive and varied ( Figure 1). Previous studies have grouped urban drivers in marine systems into three primary categories, with potentially nonexclusive effects on microevolution: pollution, ocean sprawl, and resource exploitation (e.g., Todd et al., 2019). Pollution from point and nonpoint sources can cause contamination of sediment and water and can exert strong selective pressures on populations. Ocean sprawl encompasses the construction of artificial structures, which often represent a complete habitat replacement compared to the soft sediment that they are frequently built on, and as such can massively alter ecological communities and connectivity (e.g., Bishop et al., 2017;Firth, Browne, Knights, Hawkins, & Nash, 2016), as well as genetic diversity and structure across populations (Henry et al., 2018). Resource exploitation, which includes fisheries (both commercial and artisanal/recreational), mariculture, sediment dredging, mineral extraction, and oil and gas development, might not appear as an important urban phenomenon at first glance, but are these activities are often much more intensive near urbanizing coastal ecosystems, and especially in the case of fisheries or shellfish harvesting, represent earlier historical impacts (Kirby, 2004). Mosaics of these stressors are typically found near urbanized coastlines, resulting in heterogeneous selective regimes across relatively small spatial scales (Breitburg, Hondorp, Davias, & Diaz, 2009). In this synthesis, we maintain these categories as a framework and expand on their potential and documented effects on microevolution. We also consider how interactions between particular impacts may act synergistically to shape evolutionary processes in marine organisms.
In general, marine species span the spectrum of life-history characteristics that affect how rapidly they are able to evolve, including generation time, reproductive mode, and effective population size (e.g., Ansell, Gibson, Barnes, & Press, 1994;Ayre, Minchinton, & Perrin, 2009;Hedgecock, 1994). Many marine invertebrates and some fishes have relatively short generation times, allowing populations to adapt more quickly to changing environmental conditions. However, effective population size is often much lower than census population size in marine species due to sweepstakes reproductive success and other factors (Beaumont, 1994;Hedrick, 2005;Kalinowski & Waples, 2002;Waples, 2016), suggesting available genetic variation may not always be as great as it might appear.

| Marine versus terrestrial evolution in urban ecosystems
In considering a framework for urban marine evolution, it is useful to consider how evolutionary processes in city-adjacent marine systems may mirror or differ from those in urban terrestrial systems, where such processes have been better studied. Previous studies examining evolution in urban terrestrial species indicate that many populations experience accelerated genetic drift and reduced gene flow, as well as divergent selection in life-history traits between urban and nonurban environments (reviewed in Johnson & Munshi-South, 2017;Rivkin et al., 2019). These patterns are driven by habitat loss and/or fragmentation, presence of impervious surfaces, pollution (including air, noise, water, and light), and higher temperatures (Rivkin et al., 2019). Some studies also suggest that despite differences between cities, urbanization itself sometimes creates convergent environments that may shape similar evolutionary trends across independent instances (Groffman et al., 2014;McKinney, 2006).
While urban marine evolution shares many of the same features, patterns in marine systems may not conform to terrestrial trends entirely. Parallel impacts include habitat loss and alteration (including increases in impervious or hardened substrates); light and noise pollution; and ecological effects of invasive species. It is likely that in marine, as in terrestrial, habitats, good dispersers are less likely to lose genetic diversity due to urban impacts (Bierwagen, 2007).
However, population-level responses in urban seascapes may differ from terrestrial counterparts due to contrasts in both the nature of stressors (including duration and spatial extent) and organismal characteristics, such as higher frequency of planktonic life stages and extended propagule dispersal. Impervious or hardened substrates in marine systems may be used by organisms that colonize rocky reef habitats and may even increase dispersal via stepping-stone effects (though these habitats are often not ecologically equivalent to natural rocky shorelines (see "Ocean Sprawl" below)). In addition, some of the stressors faced by marine populations (oxidative stress, harvesting pressure) do not have direct terrestrial urban equivalents.
The central goals of this review are to summarize the evidence for microevolutionary processes in urbanized marine ecosystems from previous studies and to identify important questions and areas for future research. We have limited our review to brackish, estuarine, and marine species, as strictly freshwater species differ in characteristics such as connectivity and dispersal and often experience a different set of anthropogenic pressures. We have also limited our analysis to metazoans, as the mechanisms by which marine prokaryotic communities evolve and adapt to urban stressors also differ substantially and have been covered in other reviews (e.g., Imran, Das, & Naik, 2019). While our primary focus is on documented evolution of marine populations via changes in allele frequencies, we also make note of studies that provide secondary lines of evidence that evolution may be occurring, such as common garden experiments. We find there is excellent potential for future research efforts to build on these studies using new genomic technologies, in order to fully explore evolutionary dynamics in urbanized seascapes.

| Ocean sprawl
Ocean sprawl encompasses artificial or engineered structures in marine or coastal environments, including shipping infrastructure (berths, moorings, shipping canals), protective structures (jetties, F I G U R E 1 Mosaic of urban impacts on evolutionary processes including pollution, ocean sprawl, and extractive activities. These impacts can vary widely across space (from localized to widespread) and time (from short to long time frames) sea walls, breakwaters), and residential or commercial waterfront construction (bridges, canal estates, waterfront condominiums). In addition, ocean sprawl includes structures associated with marine extraction (wind farms, aquaculture) that may be located farther offshore. While there is no total estimate of the global extent of ocean sprawl, several trends indicate that most of these categories are likely to increase rather than decrease. For example, rising sea levels and the increasing frequency of coastal storms due to climate change (Hinkel et al., 2014;Nicholls & Cazenave, 2010) have prompted many coastal cities to further modify and fortify shorelines through hardened structures. The shipping industry continues to dominate global transport of goods (Ehler, Zaucha, & Gee, 2019) resulting in a growing network of ports and associated infrastructure to protect these ports. As a result, >50% of shorelines have been hardened in some parts of Asia, Europe, and America (Bacchiocchi & Airoldi, 2003;Bulleri & Chapman, 2010;Lee & Li, 2013), including some of the largest cities such as New York, Hong Kong, and Sydney (Chapman & Bulleri, 2003;Gittman et al., 2015;Lam, Huang, & Chan, 2009). Because human settlements have often preferentially been built near rivers globally (Fang et al., 2018), many urban centers are constructed near river mouths or estuaries in which sedimentary bottom habitats predominate. The transformation of the coast to hard structures often has strong effects on biological communities, especially benthic organisms (Heery et al., 2017). Artificial structures can change water flow regimes, affecting the movement of larval propagules and potentially affecting water temperature, oxygen, and light exposure Gittman et al., 2015).
Ocean sprawl can affect marine evolution via several mechanisms depending on the nature of the artificial environments created, but perhaps the best-studied effects are on genetic diversity and connectivity. Bishop et al. (2017) review the effects of ocean sprawl on ecological connectivity between marine communities, demonstrating that artificial structures can act as both barriers to and conduits for movement of species and propagules. Because of variable dispersal potential across species, the mosaic landscapes created by ocean sprawl can fragment populations for soft-bottom taxa, as there is increasing evidence that many benthic invertebrate may disperse at small scales (e.g., Costantini, Carlesi, & Abbiati, 2013;Darling, Kuenzi, & Reitzel, 2009). Reduced gene flow near hardened structures such as sea walls has been documented in the limpet Patella caerula on the coast of Italy (Fauvelot, Bertozzi, Costantini, Airoldi, & Abbiati, 2009), as well as in soft corals in the Gulf of Mexico (Atchison, Sammarco, & Brazeau, 2008;Sammarco, Brazeau, McKoin, & Strychar, 2017). However, artificial structures can also create a stepping-stone effect for hard-substrate fauna, creating gene flow between habitats that were previously unconnected Fowler et al., 2019). This stepping-stone model suggests that along urbanized coastlines, species that specialize on hard substrates would be able to move between natural habitats more rapidly due to the construction of artificial hard substrate in urban coastal areas. Several studies on oil and gas platforms have shown low levels of differentiation between structures (corals; Sammarco, Brazeau, & Sinclair, 2012;polychaetes, Fauvelot, Costantini, Virgilio, & Abbiati, 2012). Coolen et al. (2020) combined particle tracking models with genetic data to show that populations of the blue mussel Mytilus edulis were able to use offshore oil and gas installations as stepping stones to expand their range. However, they found high genetic differentiation among some of the structures, indicating populations were colonized by few individuals (founder effects); moreover, discrepancies between modeling and genetic results suggested that patterns of differentiation were shaped both by average migration patterns and by rare events such as storms. While oil and gas platforms are not themselves structures that are typically associated with urbanization, they bear similarities to many urban structures in that they represent hardened, artificial habitat replacing soft-bottomed habitat, and therefore, studies of these habitats may yield useful insights for predicting evolutionary patterns in urban environments.
Artificial structures also create new selective regimes for organisms whose natural habitat might be rocky reefs or shores, as they differ from these habitats in terms of surface complexity, material and rugosity, movement, and disturbance frequency and extent (Airoldi & Bulleri, 2011;Bulleri & Chapman, 2004Chapman & Blockley, 2009;Moreira, Chapman, & Underwood, 2006;Tyrrell & Byers, 2007); moreover, as they are frequently colonized by opportunistic and non-native species, they can introduce new predators, food sources, or parasites (Airoldi, Turon, Perkol-Finkel, & Rius, 2015;Bracewell, Robinson, Firth, & Knights, 2013;Connell, 2001;Dafforn, Johnston, & Glasby, 2009;Firth et al., 2016;Glasby, Connell, Holloway, & Hewitt, 2007). Moreira et al. (2006) found that despite the fact that Siphonaria denticulata (pulmonate limpets) are abundant on sea walls in Sydney harbor, individuals are small in size and the reproductive output of the population is lower, suggesting this artificial habitat may be selecting for these features over evolutionary time scales. A study of hydroid assemblages in harbor sea walls versus natural rocky cliffs demonstrated significant differences between the two habitats, showing that harbor species were more likely to be small, short-lived, and with a free-swimming medusa phase-in other words, the harbor environment appears to select for r-strategists (Megina, González-Duarte, López-González, & Piraino, 2013). Moreover, Turon, Tarjuelo, Duran, and Pascual (2003) showed that populations of the ascidian Clavelina lepadiformis living in harbors in the Mediterranean were more closely related to Atlantic coast populations than to those living outside harbors on rocky cliffs, raising the possibility that this cryptic invasion was facilitated by preadaptation to turbid and polluted conditions. Artificial structures also change water flow, concentrating pollutants such as metals and nutrients and creating additional selective pressures. For example, swing moorings in Sydney Harbor have been shown to cause localized areas of concentrated heavy metals and differences in sediment grain size, with consequences for benthic and meiofaunal species (Hedge, Dafforn, Simpson, & Johnston, 2017).
Coastal hardening can create barriers between terrestrial and marine habitats for both species who move between the two (e.g., sea turtles, terrapins, pinnipeds) and sandy beach invertebrates such as TA B L E 1 Studies demonstrating changes in genetic diversity or structure associated with urban pressures in marine or estuarine species

| Pollution
Urbanization of coastal areas is nearly always associated with increased pollution in marine ecosystems. Numerous studies have   Tables 1 and 2). In addition, long-term pollution can cause large-scale changes in biological communities from the bottom to the top of food webs. For example, contaminants of all classes were found in a meta-analysis to reduce both richness and evenness in marine communities (Johnston & Roberts, 2009 Dachs & Méjanelle, 2010;Kennish, 1997;Tayeb, Chellali, Hamou, & Debbah, 2015). Many of these chemical pollutants bioaccumulate, affecting reproduction and basic cellular processes.
Sewage runoff can cause both nutrient pollution and high inputs of pathogens and parasites. In many urban settings in North America and Europe, runoff from combined sewer overflow pipes (CSOs), in which storm water and human sewage are merged and discharged into waterways, is a particularly acute issue that can result in significant ecosystem impacts. These effects include long-term eutrophication, oxygen depletion and ammonia peaks (Borchardt & Sperling, 1997), localized acidification (Howarth et al., 2011), increases in PAHs (Launay, Dittmer, & Steinmetz, 2016), organochlorines and wastewater micropollutants (e.g., estrogens, caffeine, ibuprofen) (Ellis, 2006;Phillips et al., 2012), and exposure to human and animal pathogens (Calderón, Porter-Morgan, Jacob, & Elkins, 2017). CSOs have been shown to impact the structure and diversity of benthic communities (Seager & Abrahams, 1990) in some cases, whereas no effects have been identified in others (Rochfort et al., 2000). Pollution from outfalls can hinder larval dispersal and thus gene flow by creating localized environments that are toxic or deadly to larvae (Puritz & Toonen, 2011) or may cause local directional selection for certain genotypes (Larsson et al., 2016). While no previous studies appear to have examined adaptation of metazoan communities near CSOs, these sources of contaminants and nutrients are likely to select for organisms that are able to tolerate a wide range of chemical stressors. Benthic, stationary species such as mollusks are most likely to be affected by CSOs as they will experience the most direct effects; however, mobile species may also be impacted by increases in pathogens, especially where the zone of CSO impacts overlaps with feeding and breeding grounds. For example, sea otters and pinnipeds have been shown to be susceptible to parasites and diseases from urban runoff (Miller et al., 2010;Stoddard et al., 2008).
Beyond industrial or chemical toxicants, urban pollution of marine ecosystems extends to many other sources that are less well-stud- Evolved resistance to pollutants is the most commonly documented mode of evolution and has been studied extensively in several marine species (Table 2), with most studies focused on effects of heavy metals and, to a lesser extent, organic pollutants. A full review of evolved resistance to pollutants by marine animals is beyond the scope of this article but is addressed in many previous reviews teri evolved resistance to cadmium and nickel at a highly polluted estuary site in the Hudson River within 3-4 generations (Klerks & Levinton, 1989). Follow-up studies then documented the loss of resistance following the cleanup of the site (Levinton et al., 2003).

| Resource extraction
While resource extraction-including fisheries and aquaculture-is perhaps less obvious as a driver of evolution in urbanized marine environments compared with ocean sprawl and pollutants, extractive activities nearly always predate the other two categories in the history of any urban settlement and generally represent the oldest anthropogenic stressors on marine systems resulting from high human density along coastlines (Jackson et al., 2001). In many cases, past exploitation of marine populations near human settlements reshaped entire ecosystems early in their histories; for example, overharvesting of fisheries on the east coast of the United States began near urban centers such as Boston and sequentially depleted neighboring areas (Kirby, 2004 Bulleri, & Abbiati, 2005). In addition to these trends in fisheries, urbanization corresponds with growing extractive industries including shoreline or nearshore dredging for sand, minerals, rock, and coral for construction or other purposes (Charlier & Charlier, 1992 Post et al., 2002). This suggests that these activities do have the potential to affect microevolution via reductions in genetic diversity and fisheries-induced selection (Lewin, Arlinghaus, & Mehner, 2006). For example, in fishes with genetically determined seasonal migration, as in salmon, selective angling in certain seasons may interfere with adaptive structuring of populations (Consuegra, Leániz, Serdio, & Verspoor, 2005).
In addition to these three categories,  (Mullen, Peterson, & Todd, 2013). This species has been called "the urban whale" due to the overlap in its habitat with human activities (Kraus, Rolland, & Rolland, 2007).
While the long generation time of this species likely precludes genetic adaptation to urbanization, erosion of genetic diversity due to mortality from ship strikes and entanglement near cities may constrain future adaptation to changing global conditions. Finally, in some urban areas, food provisioning by humans may be changing parasite load, gut microbiome, and timing of reproductive behaviors, with implications for long-term adaptation, as has been shown in wading birds in southern Florida wetlands (Kidd-Weaver et al., 2020;Murray et al., 2020).

| Interactions between stressors
The effects of anthropogenic stressors on microevolutionary dynamics cannot be considered in isolation as they may act cumulatively-additively, synergistically, or antagonistically-on marine populations (Crain, Kroeker, & Halpern, 2008;Todd et al., 2019).
Results from previous meta-analyses suggest that multiple stressors often act synergistically in marine systems, in contrast to terrestrial and freshwater ecosystems (Crain et al., 2008;Harvey, Gwynn-Jones, & Moore, 2013). Different stressors may overlap in time, or may occur sequentially as cities grow and develop. Many stressors, particularly those that result in large-scale habitat loss such as ocean sprawl, are likely to reduce population size substantially, which can erode standing genetic variation and hinder the ability of populations to rapidly adapt to acute stressors such as pollution. Our synthesis indicates that a wide variety of urban stressors from artificial structures to pollutants are associated with declines in genetic diversity in diverse marine species (Table 1); reduced genetic diversity may compromise the ability of these populations to adapt to other stressors.
In addition to loss of genetic diversity, populations living on artificial structures may be smaller-bodied than populations living on natural hard surfaces (Megina et al., 2013;Moreira et al., 2006).  (Piola & Johnston, 2008).
As an example of antagonistic interactions, nutrient enrichment can lessen the harmful effects of toxins in some cases (Breitburg et al., 1998). On the other hand, prevalence of infection by parasites was found to increase in oysters that were exposed to other stressors like higher sediment load or decreased water flow (Lenihan, Micheli, Shelton, & Peterson, 1999).
Perhaps the most important cumulative effects on evolution in urban seascapes, and yet among the least studied to date, will come from the physiological challenges to marine organisms presented by climate change and ocean acidification, when combined with ongoing urban stressors. Clear examples of these complex problems come from the synergistic effects of nutrient enrichment and other stressors such as sea level rise and predator loss on estuarine salt grasses (Deegan et al., 2007(Deegan et al., , 2012. Warming waters can also lead to invasion of parasites or competitors (Stachowicz et al., 2002).
Generalized stress from climate change impacts to water temperature, pH, and salinity may also reduce the ability to resist and recover from exposure to pollutants (Moe et al., 2013).
Stressors that occur at the same time can present challenges to individuals along multiple physiological axes, and timing of stressors can affect physiological and potentially evolutionary outcomes (Bible et al., 2017;Molinos & Donohue, 2010). Stressors that act in concert (simultaneously) can potentially delay or change evolutionary dynamics; for example, a study that examined the evolutionary response of bacteria to both predators and antibiotic exposure showed that the evolutionary responses to both stressors were delayed (Hiltunen et al., 2018). However, consecutive stressors can, in some cases, have larger impacts on individuals and populations than simultaneous stressors, as in the case of acidified conditions and salinity in crab larvae (Miller et al., 2014) and consecutive warmer air and water temperatures in sea stars (Pincebourde, Sanford, Casas, & Helmuth, 2012), if exposure to the first stressor causes physiological changes that leave an organism vulnerable to a second stressor (e.g., Todgham & Stillman, 2013). Jayasundara et al. (2017) showed that killifish adapted to polluted waterways demonstrated altered metabolic and physiological responses to other stressors such as temperature, indicating that adaptation to pollution may compromise the ability of this population to acclimate to other stressors. In addition, each stressor may be inconsistent and may fluctuate over time. Because no single phenotype may be optimal in the context of these multiple, fluctuating stressors, such combinations may select for broad organismal plasticity over time (e.g., Gomez-Mestre & Jovani, 2013;Hallsson & Björklund, 2012).

| C A S E S TUD IE S: R APID ADAP TATI ON IN URBAN MARINE ECOSYS TEMS
For most marine species, the speed of urban environmental change will outpace their capacity to evolve: Rapid adaptation to urban conditions is likely to be the exception rather than the rule. However, certain marine species appear to survive and even thrive in urbanized seascapes, perhaps the marine equivalents of terrestrial synanthropic species such as pigeons and raccoons. For example, the photosynthetic capacity of the green algae Ulva lactuca increases in environments with greater pollution, while that of the brown algae Sargassum stenophyllum decreases (Scherner, Barufi, & Horta, 2012).
Overall, many of the species that appear to thrive in urbanized seascapes appear to have a generalized tolerance of environmental variability, which may derive in part from preurban adaptation to such variability in native estuaries: Species that are able to persist in estuarine and other tidal environments must necessarily cope with large fluctuations in variables like temperature, oxygen, and sediment load and thus show cross-tolerance across stressors (e.g., Ravaschiere et al., 2017;Schulte, 2007;Todgham, Schulte, & Iwama, 2005). The evolution of physiological plasticity is key to allowing species to persist in challenging environments (e.g., Brennan, Galvez, & Whitehead, 2015), and thus, these species may in some cases be preadapted to stressful and variable urban conditions. In addition, adequate genetic diversity is necessary for evolutionary responses to a wide range of urban stressors. A number of studies of rapid evolution to urban stressors indicate organisms are able to adapt quickly due to standing genetic variation (see Reid et al. 2017;Wirgin et al., 2011 below), or even more intriguingly, via introgression from other species (Oziolor et al., 2019). Additional experimental evidence indicates the importance of standing variation in adaptation; for example, Nevo, Noy, Lavie, Beiles, and Muchtar (1986) showed that in several pairs of species with differing levels of genetic diversity, the species with greater diversity was found to be more resistant to toxicants. The addition of genomic data from a wider array of species may reveal more about the most common sources of genetic variation during episodes of rapid adaptation.

| Case study: Copper tolerance in macroalgae and invertebrates
Copper is one of the most common pollutants found in industrial sediments and is among the most toxic metals for both macroalgae and invertebrates, as it can inhibit photosynthesis and disrupt important cellular functions (Connan & Stengel, 2011;Gledhill, Nimmo, Hill, & Brown, 1997;Qian et al., 2009). For this reason, it is often introduced purposefully into the marine environment as an anti-foulant.
Several species of macroalgae show population and species-level differences in tolerance to copper. For example, Han, Kang, Park, Lee, and Brown (2008) showed that growth and photosynthetic rate were largely unchanged in response to copper exposure in one species of Ulva (U. armoricana), whereas both parameters decreased significantly in another species (U. pertusa). Nielsen, Brownlee, Coelho, and Brown (2003) found that populations of the brown alga Fucus serratus from copper-contaminated sites were more resistant to copper exposure, as measured by tissue concentration and growth rates, suggesting an exclusion mechanism. Moreover, similar responses in progeny of the adults indicate that the resistant phenotype is at least partially inherited. In contrast, Brown, Newman, and Han (2012) tested for interpopulation differences in copper tolerances in the brown seaweed Gracilariopsis longissimi and found no significant differences between populations from habitats differing in copper concentration, suggesting that observed copper tolerance is a species-wide feature rather than an evolutionary response. Polyploidy and the ability to hybridize may be important means by which some species of algae maintain the genetic variability needed to cope with new habitats (Coyer et al., 2006). While relatively few studies have tested genetically based and heritable tolerance in macroalgae, the variability in tolerance across populations and species indicates this system has excellent potential for future studies.
Copper is also especially toxic to many marine invertebrates such as cnidarians, mollusks, bryozoans, arthropods, and poriferans (Johnston & Keough, 2000;Reichelt-Brushett & Harrison, 2000). In the bryozoan Watersipora subtorquata, heritable differences in tolerance to copper were found across individual colonies, suggesting this species has the potential to evolve rapidly (McKenzie et al., 2011). Larval size, which has a heritable component but also results from the interaction between parental genotypes and environmental conditions, was also found to positively affect tolerance in this species. Larval size has also been shown to increase in response to pollutant levels in another bryozoan, Buguta neritia (Marshall, 2008), indicating this trait may be a plastic response to fluctuating levels of pollutants. Galletly et al. (2007) Piola & Johnston, 2006a;Sáez et al., 2015), or across different species (Piola & Johnston, 2006b, 2009) but have not explicitly tested for heritability of these traits.

| Case study: Atlantic killifishes (Fundulus heteroclitus)
The Atlantic killifish Fundulus heteroclitus is perhaps the best-studied example of rapid and repeated adaptation to pollutants. Populations of this species, which is highly abundant along the eastern seaboard of the United States and Canada, have independently evolved heritable tolerance to a variety of pollutants and toxicants, including heavy metals, PAHs, and PCBs (Nacci et al., 1999;Nacci, Champlin, & Jayaraman, 2010;Reid et al., 2017;Weis, 2002;Whitehead et al., 2017,). Extensive genomic and transcriptomic data for this species reveal a complex set of genes involved in tolerance pathways, including but not limited to the aryl hydrocarbon receptorbased signaling pathway (Reid et al., 2016

| S PATIAL SC ALE S OF E VOLUTI ON: CONS IDER ATIONS FOR URBANIZED S E A SC APE S
As is the case in terrestrial urban ecosystems, the temporal and and exposure can vary with season due to interactions between pollutants and organic matter (Birch, Taylor, & Matthai, 2001;Buggy & Tobin, 2008). In contrast with this localized scale of stressors, many species of nearshore invertebrates and fishes have a pelagic larval stage that ranges from days to months, and a sedentary benthic adult stage, suggesting a much wider spatial scale of dispersal. This mismatch between the small scale of the selective pressure and the large potential range of gene flow suggests that adaptive processes may often be slowed in urban coastal species; indeed, there is evidence for this idea from some comparative studies across similar species with different larval dispersal (Untersee & Pechenik, 2007).
However, there are many exceptions to this general model: For example, many stressors have wider-ranging impacts within and beyond the urban coastline (shipping noise, coastal construction), and a number of marine species including some fishes, gastropods, and crustaceans have direct development or very short larval duration rather than extended larval dispersal, generally with expected consequences for genetic structure (e.g., Collin, 2001;Dawson, 2001;Kyle & Boulding, 2000). Moreover, both natural (oceanographic) and human-made barriers to dispersal may result in gene flow occurring across smaller scales than larval duration would suggest (Marko 2004;Marko, Rogers-Bennett, & Dennis, 2007;Pelc, Warner, & Gaines, 2009). For species that produce larvae in patchy or enclosed habitat (e.g., estuaries or lagoons), gene flow may be further reduced below expected levels (Watts & Johnson, 2004). Indeed, studies in Table 1 indicate that urbanization is often correlated with decreases in genetic diversity across a wide variety of species with different reproductive modes, suggesting that gene flow between urban and reference sites is not extensive enough to homogenize alleles.
To further complicate matters, coastal cities may themselves also generate new opportunities for dispersal between and out of urban marine ecosystems, both via creation of artificial stepping stones for hard surface dwelling organisms (see Ocean Sprawl above) and also by generating new opportunities for transport via ballast water and rafting on marine debris or tar balls. Sessile calcareous organisms such as serpulid polychaetes, bryozoans, barnacles, and gastropods are often found encrusted on floating plastic debris, and this mode of transportation has allowed, in some cases, long-range dispersal away from native habitat (e.g., Winston, 2012) and promotes invasion by nonindigenous species (Campbell et al., 2017).
While gene flow from nearby unaffected ecosystems can slow the process of adaptation by homogenizing populations, in cases where survivorship in an urban-impacted system is low, those source populations may harbor crucial genetic diversity that allows adaptation to occur. Thus, the proximity of urbanized seascapes to nearby intact ecosystems should be considered when studying the speed of evolution and in planning protected areas.

| CON CLUS I ON S , OUTLOOK , AND RECOMMENDATIONS FOR FUTURE S TUD IE S
As human density continues to rise near coastlines, activities ranging from ocean sprawl to pollution will exert increasing selective pressures on nearshore organisms living in urbanized seascapes. In many urban marine ecosystems, selective regimes may change quickly over time, and/or may be heterogeneous over small spatial scales, to the extent that the most evolutionarily successful species will be those with large standing genetic variation and those with high phenotypic plasticity. (genetic diversity and structure) were carried out using very few or anonymous markers (Table 1). Newer methods such as RADseq allow the rapid assessment of genetic diversity and structure at a genomewide scale, permitting a much more thorough analysis than was previously possible and opening the possibility of more finescaled demographic analyses of recent bottlenecks; however, it is important to note that RADseq methods require sufficient marker density across the genome in order to have utility in identifying adaptive regions (e.g., Catchen et al., 2017;Lowry et al., 2017a;Lowry et al., 2017b). For the purposes of studying adaptation, methods such as low-coverage whole-genome sequencing, Pool-seq, exome capture, and RNAseq are superior to RADseq and could be used to expand on experimental studies such as Nevo et al. (1986) to compare tolerance and adaptive responses of populations or species with differing standing genetic variation (Hoban et al., 2016;Jones, & Good, 2016;Puritz & Lotterhos, 2018;Schlötterer, Tobler, Kofler, & Nolte, 2014).
Another valuable direction for future research would be use of historic specimens (e.g., archived scales) to analyze genomic changes over many generations and to assess hypotheses about the origin of adaptive alleles. Here again, new methodologies such as Hy-RAD and targeted capture have improved the chances of sequencing DNA from old specimens. Of particular interest would be urban marine populations near older cities (>400 years), where evolution has had more time to occur, such as Lagos, Osaka, Jakarta, Panama City, Singapore, New York, Boston, and Sydney. However, a challenge with older cities is that, due to greater urban sprawl, it is often more difficult to find nearby and relevant reference (nonurbanized) sites for comparison.
With sea level rise, higher temperatures and increases in extreme weather events, urban dwellers, and policymakers are increasingly aware of the importance of ensuring the resilience of coastal marine ecosystems to protect coastlines and restore biodiversity (Dafforn et al., 2015;Moschella et al., 2005;Munsch, Cordell, & Toft, 2017;Perkins et al., 2015;Strain et al., 2018). However, evolutionary potential is rarely included as an explicit factor in planning marine habitat restoration and interventions. To allow thriving marine communities on the edge of dense cities, it is critical that urban planners and designers consider evolutionary processes. Shoreline development can be mandated to include habitat improvements that will help stabilize marine population sizes and therefore help maintain population genetic diversity-for example, in a meta-analysis of 109 studies, Strain et al. (2018) found that the majority of eco-engineering interventions in urban marine ecosystems increased the abundance of particular species and/or diversity of species. Promising directions include a new residential development along New York City's East River that includes restored oyster reefs, salt grass, and small tidal inlets. Such "blue spaces" also present great opportunities for education and community engagement to help city dwellers learn about and eventually advocate for protecting urban marine habitats (Dafforn et al., 2015). In addition to encouraging habitat improvements as part of construction projects, any development of artificial structures near cities should include consideration of how these structures may hinder or facilitate gene flow. As marine urbanization increases over the coming decades, improving our knowledge of how marine species are evolving in these altered habitats, as well as working to conserve them, will be critical to minimizing the longterm impacts of urban pressures on bordering seascapes.

ACK N OWLED G M ENTS
We are grateful to John Waldman, Nathan Morris, Vimal Ramnarain, Davendra Ramnarain, Zahra Alkaifi, and Rose Kaluigwe for helpful discussions, as well as funding from the National Science Foundation (Award #1433014).

CO N FLI C T O F I NTE R E S T
None declared.