Annual recruitment of Hymeniacidon perlevis on fouling plates was monitored across six primary sites from 2007 to 2023, with 4 additional sites surveyed from 2010–2014. For complete details of how plates were deployed and processed, see Wasson et al. (2020). In short, five replicate ceramic plates (10 × 10 cm) (Suppl. material 1: fig. S1) were deployed across ten sites and replaced in May or June annually (Suppl. material 1: fig. S1, Table 1), with the primary purpose of detecting native oyster recruitment. Percent cover of H. perlevis was recorded on each tile deployed except for 2018–2020, when only presence/absence was recorded. Sponge cover data were averaged across all 5 plates for a site.

Visual surveys (quadrat) of sponge tissue were recorded at three sites in the upper slough monthly for 2 years from August 2021 to July 2023. Sites were selected based on their proximity to an NERR water monitoring instrumentation array, accessibility by foot, substrate variability, and previously recorded observations of Hymeniacidon perlevis (Fig. 1C). Surveys were undertaken by dividing each site into 1 m wide transects, then photographing a 1 × 1 m quadrat at different intervals along each transect. Transect number, length, and quadrat intervals were determined by the shape and accessibility of each site. Photographs of each quadrat were taken with a smartphone and sponge area within each image was measured by tracing polygons using Fiji ImageJ software (Schindelin et al. 2012) (Suppl. material 1: fig. S2). Monthly mean sponge cover was calculated by randomly selecting 38 quadrats from each site to standardize results (38 was the minimum number of quadrats measured across all sites and months). To assess whether sponge cover varied over time, a bootstrapped one-way ANOVA was calculated using month as a factor. We also opportunistically collected 10 sponges each month during a period of peak cover from three sites to determine whether sponges were reproductive.

Figure 1. 

Scale of H. perlevis introduction in Elkhorn Slough. A. A large aggregation in South Marsh; B. A large individual exposed at low tide; C. Map of Elkhorn Slough, CA with monthly monitoring sites marked with stars and long-term monitoring sites marked with dots. APN: Azevedo Pond; KP: Kirby Park; NM: North Marsh; WSL: Whistlestop Lagoon; RSM: South Marsh – WSL; RBR: South Marsh Bridge; VM: Vierra Mouth; BSE: Bennett Slough; MLN: Moss Landing Road; JR: Jetty Road S1 and S2 mark the NERR sondes which collected water quality parameter data; D. Sponge cover was measured using 1 × 1 meter photoquadrats. Map: Created in ArcGIS Pro; service layer credit Esri, FAO, NOAA, USGS, California State Parks, Esri, HERE, Garmin, SafeGraph, METI/NASA, USGS, Bureau of Land Management, EPA, NPS, USDA.

The Azevedo Pond site (APN) is in a channel between a culvert accessing a small, shallow tidal pond and the upper portion of the Elkhorn Slough main channel. The substrate is composed primarily of cobbles but is broken by patches of larger stones or mud. Tidal changes force water swiftly through this channel due to the depth difference between the main channel and the pond. Sponges here were found in dense clusters populating the cobbles and large stones. At APN, eight adjacent transects took up the width of the channel and were deployed out 20 m, at which point the water became too deep as it entered the main channel. Five quadrats were photographed along each transect at randomly generated points.

The Kirby Park site (KP) lies on a tidal mud flat adjacent to the upper portion of the Elkhorn Slough main channel. The mudflat is broken up with the occasional boulder and the southern edge is bounded by a long horizontal cement piling. Sponges at this site were spread among the boulders and some small stones across the sediment and had a tendency to be partially buried. At KP, ten transects were randomly placed between 0 and 30 meters north of the cement piling, and deployed out 5 m, at which point the mud became too deep. Five quadrats were photographed along each transect at randomly generated points.

The South Marsh Bridge site (RBR) lies adjacent to an upraised gravel roadway held in place by wooden walls. One edge of the site is at the mouth of a levee that experiences rapid currents when the tide changes. Sponges at this site were found in a narrow band, approximately 3 m wide, as clusters and occasional sheets on the cobbles and adjacent rip-rap. A few individuals colonized the wooden walls of the levy as well. At RBR, thirty adjacent transects were placed between 0 and 30 meters north of the bridge and were deployed to 3 m, at which point the shallow, muddy water made sponge photographs difficult. Three quadrats were photographed along each transect to compensate for the short transect length.

We assessed whether sponges were reproductive during their period of peak cover by surveying for brooded embryos monthly from September through December 2022. Severe storms in January and February 2023 ended sampling. We sampled from three sites (distinct from the sites where quadrats were surveyed) monthly: one north of the Kirby Park site, and one on each side of the Azevedo Pond culvert (near to but outside of the quadrat sampling area). Ten individuals were dissected in the field and assessed visually to determine if embryos were brooded in the tissues using a 20× loupe.

Publicly available National Estuarine Research Reserve (NERR) monitoring data collected from two YSI EXO2 multiparameter instruments (Fig. 1C) in the Elkhorn Slough were downloaded from the Centralized Data Management Office (https://cdmo.baruch.sc.edu/). Monthly average air temperature (°C) and total monthly precipitation (mm) were recorded from Caspian Weather Station. Monthly (day 01–31) averages for water temperature (°C), acidity (pH), salinity (ppt), dissolved oxygen (mg/L), turbidity (NTU), and precipitation (mm) were calculated from continuous data collected every 15 minutes. Correlations between sponge cover and environmental data were tested using Spearman rank cross-correlations. Since changes in sponge cover may be expected to lag behind environmental changes, time-lagged correlations were also assessed, with monthly sponge cover delayed behind environmental conditions by 1 to 6 months. Monthly averages for environmental variables were calculated using the stats package in R version 4.2.2 (R Core team, 2022). The strongest positive or negative correlation coefficient (rho) for any given environmental parameter (and its time lag) was considered to provide the greatest evidence of correlation with sponge cover.

To estimate the total volumetric flow from all sponges at a given site, cover was related to pumping rates measured from individual sponge oscula in the field. First, mean osculum density on a sponge was calculated by photographing sponges at APN top-down and counting the number of oscula within a randomly selected cm² of sponge (n = 22). Average osculum size (cm²) was measured by randomly selecting a single osculum from each cm² and measuring its diameter using Fiji ImageJ (n = 22). Half of this average diameter was used to calculate the average osculum area assuming a circular osculum. Error propagation was calculated from products using the formula (Taylor 1997):

$\text{ If }x=a\times b, \Delta x=x\sqrt{{\left(\frac{\Delta a}{a}\right)}^2+{\left(\frac{\Delta b}{b}\right)}^2}$

The velocity of excurrent flow was calculated using the dye front technique following Yahel et al. (2005). In brief, a graduated plastic cylinder was filled with a small amount of fluorescein dye at one end and capped with a finger at the other end while underwater. The cylinder was placed in front of an osculum without touching it and the finger was then removed. The movement of the dye was recorded with a GoPro Hero9 camera and the time for the dye front to reach the end of the cylinder was recorded using Kinovea video annotation software (www.kinovea.org). The velocity of the dye front was averaged across 3 sponges with 1 osculum assessed from one sponge, 2 from the next, and 7 from the third. Each osculum was sampled between 1 and 7 times (n = 10 total oscula). This average was multiplied by the average osculum area to estimate the average volumetric flow assuming plug flow.

We related two-dimensional sponge cover to estimate biomass using morphometrics on a subset of sponges. First, full individuals were removed from rocks and photographed top-down to calculate their two-dimensional area in ImageJ. Volume of those same individuals was measured using water displacement. A linear regression with an intercept of zero was used to determine the relationship of sponge area:volume using the regression slope (n = 15). (Suppl. material 1: fig. S3). A similar relationship was calculated between sponge volume, wet weight, dry weight, and ash-free dry weight. Displacement volumes of wet sponges collected from the field (n = 7) were recorded, then sponges were rinsed in fresh water and wet weights were recorded in pre-combusted aluminum trays. Tissue samples were subsequently dried overnight at 55 °C and weighed again to determine dry weight. Finally, the same samples were placed in a muffle furnace for 4 hours at 400 °C to determine ash-free dry weight. Organic weight was calculated by subtracting ash-free dry weight from dry weight. From these relationships, the average biomass per mL of tissue was estimated (Suppl. material 1: fig. S3).

To estimate biomass and volumetric flow for the sponge cover at a given site, the morphometric relationships measured above were scaled with observed cover. Volumetric flow was multiplied by the average number of oscula cm^-2 sponge cover, and further multiplied by the estimated average cover at each site each month to approximate the full extent of water processing by sponges at each site assuming continuous flow. Error propagation was calculated using the equation of Taylor (1997) above. Similarly, by multiplying the average volume:area ratio by the estimated average cover, an organic weight for sponges at each site every month was calculated.

Settlement plates used to track the spread of Hymeniacidon perlevis through Elkhorn Slough showed a generally growing and expanding population over a period of 17 years, but with variation across sites and years (Fig. 2, Suppl. material 1: fig. S4, Suppl. material 2: table S1). When initially discovered in 1998, H. perlevis was found across 200 m of shoreline parallel to the main channel. At its greatest extent, from 2019 to 2021, it was found across 3.5 km of shoreline. In 2023, this range reduced to 2.3 km of shoreline. H. perlevis was detected at six sites in the upper estuary and was not detected in any sites in the lower estuary during the sampling period. Though fouling plate deployments were discontinued in the lower estuary after 2014, there have also been no anecdotal observations from any researchers in the Elkhorn Slough reported since 2014 either, suggesting that recruitment has not occurred in significant numbers (Suppl. material 2: table S1). While H. perlevis has successfully spread upstream within Elkhorn Slough, there is no evidence it has moved seaward, with its lowermost extent remaining at the RBR site, approximately 4.7 km from the mouth of the slough. Despite high variability in detection and percent cover at various sites, annual surveys confirm successful recruitment every year from 2014 to the present.

Figure 2. 

Hymeniacidon perlevis detections from settlement plates at several sites within Elkhorn Slough over time. The shaded area represents the mean cover of sponges on annual recruitment plates at each site. Blue indicates that no plates deployed at a site were colonized that year, and green represents colonization occurred that year. White represents a period of no sampling at a site, open green boxes indicate H. perlevis presence, but no cover data were recorded. Sites are ordered from upper to lower estuary, with lower slough sites beginning at VM (Vierra Mouth).

Monthly surveys of sponge cover over two years at three sites in Elkhorn Slough revealed seasonality (Fig. 3). Generally, increases in biomass were detected beginning in the summer (August 2021, June 2022), peaked in fall, and then declined in the winter months. From March through May in both years sponges were undetectable (either absent or covered beneath sediments) or at low cover compared to other parts of the year (Fig. 3, Suppl. material 2: table S2). Embryos were detected in sponge tissues during the three months of peak biomass (September – November 2022) at all locations surveyed suggesting seasons of peak biomass correlate with active reproduction, although the full phenology in Elkhorn Slough is unknown.

Figure 3. 

Mean percent cover (± standard error) of H. perlevis from three sites over two years of monitoring. Note overlap of Azevedo Pond and Kirby Park sponge cover, as both had 0% cover from March 2023 to July 2023.

While the overall pattern of sponge cover across the year held true, sponge populations at the three sites varied in cover and timing. Sponge cover at APN (Bootstrapped ANOVA, trials = 1,000, p < 0.001) and KP (Bootstrapped ANOVA, trials = 1,000, p = 0.001) varied between months, but no significant variation in sponge cover between months was found at RBR (Bootstrapped ANOVA, trials = 1,000, p = 0.171). Despite this discrepancy, sponge cover at RBR followed similar trends to cover at APN and KP (Fig. 3, Suppl. material 2: table S2), reaching a maximum in late fall and minimum in the spring.

Monthly variations in sponge cover were found to lag behind changes in environmental conditions (Table 2). The strongest cross-correlations between several environmental variables and sponge cover at APN, KP, and RBR occurred with a 2–4 month lag, indicating that this is the time until sponge cover visually reflects changes in the environment in Elkhorn Slough (Table 2). A strong positive correlation was observed between sponge cover and temperature (measured from both water surface and atmosphere) and a negative correlation between cover and dissolved oxygen. Sponge cover at APN was significantly correlated (p < 0.05) with all parameters except for precipitation (p = 0.06). Precipitation was only found to be significantly correlated with sponge cover at KP (p = 0.02) (Table 2).

Relationships between sponge cover, biomass, volume, osculum density cm^-2 of sponge, and water pumping rate were used to estimate the volume of water filtered and sponge biomass, scaled up to Hymeniacidon perlevis cover within each site. The quantity of water processed and sponge biomass changed seasonally (Fig. 4, Suppl. material 2: table S3). H. perlevis populations from these three sites in Elkhorn Slough had on average 2.36 ± 0.28 oscula cm^-2 sponge area when viewed from above (n = 22, mean ± SE). The average size of each osculum was 0.65 ± 0.08 mm² (n = 22). The average volume pumped by an active osculum, measured by the dye front method, was 31 ± 3 mL h^-1 (n = 10).

Figure 4. 

Estimates of volume of water filtered and biomass,calculated using H. perlevis cover data within each sample site. Error bars are not shown due to excessive size.

To estimate potential variation in seasonal biomass and water filtration, we first assessed relationships between several morphometric measurements (Suppl. material 1: fig. S3). A single cm² of Hymeniacidon perlevis when measured from above had a volume of 3.4 ± 230 mL (n = 15). For 1 mL of H. perlevis tissue, the average wet weight was 0.63 ± 0.31 g (n = 7), dry weight was 0.10 ± 0.056 g (n = 7), ash-free dry weight was 0.048 ± 0.033 g (n = 7), and organic biomass was 0.054 ± 0.029 g (n = 7). Using these relationships, we estimate that sponges at APN had the greatest biomass and water filtration, based on their cover. APN sponges pumped 36 ± 9 L h^-1 m^-2 in November 2022 and 0 L h^-1 m^-2 in March 2023 (0.09 ± 13 kg m^-2 to 0 kg m^-2), KP sponges pumped 8.6 ± 2.8 L h^-1 m^-2 in November 2022 and 0 L s^-1 m^-2 in March 2023 (0.024 ± 3.5 kg m^-2 to 0 kg m^-2), and RBR sponges pumped 18 ± 8 L h^-1 m^-2 in November 2022 to 9 ± 5 L h^-1 m^-2 in March 2023 (0.05 ± 7 kg m^-2 to 0.03 ± 4 kg m^-2) (Suppl. material 2: table S3). Note the large standard error, which was driven by the very patchy distribution of sponges in their habitat, variation in area:volume ratios, and propagation of that error into these calculations.

By scaling estimates from a square meter up to the full extent of intertidal zone occupied by sponges at each of the three sites, we estimate that the volume pumped by APN sponges had the greatest change across the year, from 1.6 ± 0.40 L s^-1 at peak sponge cover in October 2022 to 0 L s^-1 when sponges were absent, in March 2023. The corresponding change in sponge biomass was from 14.5 ± 2060 kg to 0 kg. By comparison, all KP sponges filtered a peak rate of 1.1 ± 0.35 L s^-1 at peak sponge biomass of 11 ± 1560 kg in November 2022, and 0 L s^-1 at minimum sponge biomass of 0 kg by March 2023. Sponges at the RBR site pumped a peak of 0.46 ± 0.21 L s^-1 with 4.6 ± 646 kg biomass during November 2022 and in contrast with the other two sites, sponges persisted through March 2023, which meant a calculated rate of 0.24 ± 0.12 L s^-1 pumped and 2.4 ± 340 kg of sponge biomass (Suppl. material 2: table S3).

From 2008 to 2022, recruitment varied, as is typical for many species (Connell 1985; Menge and Sutherland 1987), but there was a general increase in the number of sites, frequency, and cover of Hymeniacidon perlevis recruits on settlement plates. However, recruitment was only observed in upper estuary sites. Adults of this species are also only found in the upper estuary. This distribution follows a general pattern in Elkhorn Slough, with introduced species being more diverse and more abundant in the upper estuary (Wasson et al. 2005). The upper estuary is more eutrophic than the lower estuary, with higher chlorophyll and temperature and longer residence time (Suppl. material 1: fig. S5); APN is even considered “hypereutrophic” (Hughes et al. 2011), and also has the greatest levels of H. perlevis cover. In a study near an inland fish farm, H. perlevis grew faster and more densely in the eutrophic outflow water of the farm compared to neighboring areas (Mercurio et al. 2023). H. perlevis may grow in greater densities in the upper estuary due to increased food availability from eutrophication, particularly in sites such as APN where eutrophic water is provided via daily tidal flow through narrow channels. However, abiotic, density-independent factors such as severe weather incurred broad effects across all sites. For example, after several years of consistent recruitment, sponges only recruited to a single site in 2023 (Fig. 2). This was in the wake of frequent storms with heavy rainfall in central California from November 2022 to April 2023, which generated the heaviest precipitation observed during the study period. We infer that this intense precipitation period may have in part created unsuitable conditions for H. perlevis reproduction or recruitment.

Although the populations in Elkhorn Slough follow annual cycles like other populations, the Elkhorn Slough populations reached peak above-ground biomass in mid-October and minimum biomass by February or March. In contrast, populations in the Yellow Sea and United Kingdom reach peak biomass earlier, in September (Juniper and Steele 1969; Stone 1970; Cao et al. 2007; Cao et al. 2012). The timing of minimal biomass and/or dormancy varies even more. In the Yellow Sea, sponges regressed completely from December through February; in the United Kingdom sponges declined through late spring until May but did not reach complete dormancy. In the Ionian Sea, sponges became dormant in September, which is during or even before peaks for populations in other areas of the world (Gaino et al. 2010). Within Elkhorn Slough, populations at APN and KP reached extremely low cover during late winter and spring, with most individuals becoming undetectable from above-ground and presumably regressing either completely or so that the only tissue that remained was covered by sediments. RBR populations did not disappear completely but also shrank to their smallest cover in March. In future it will be important to assess the full above- and below-ground biomass of sponges, and to determine whether this species can remain functional as a psammobiontic species or has other adaptations to a semi-infaunal lifestyle (Werding and Sanchez 1991; Ilan and Abelson 1995; Rützler 1997; Schönberg 2015).

Temperature may in part explain differences in Hymeniacidon perlevis population dynamics across various regions. Water temperatures in the English Channel (within H. perlevis’ presumed native range) only dropped to 5 °C during the winters of 1967 and 1968 when sponge cover was measured (Stone 1970), while in the Yellow Sea where sponges regress completely, the water temperature was as low as -1.1 °C in the winter (Cao et al. 2012). The minimum water surface temperature in Elkhorn Slough across this study period was 4.5 °C, much closer to the recorded temperatures in the English Channel. Air temperature at Elkhorn Slough was generally temperate but occasionally reached freezing in February and April of 2022, with -0.5 °C the lowest recorded temperature during April 2022. However, sudden population declines such as those observed in the Ionian Sea (just after temperatures peaked at 26.8 °C) (Gaino et al. 2010) indicate factors such as salinity, food availability, and dissolved oxygen influence H. perlevis’ phenology. From the 2–4 month time lags we found between sponge cover and environmental variables, we hypothesize that peak sponge cover occurred 2–4 months after conditions were typical of summer periods in Elkhorn Slough: warmer, saltier, more acidic, with lower turbidity and lower dissolved oxygen concentrations (Caffrey et al. 2002b). While the conditions at the end of winter may trigger new growth, we hypothesize that the conditions required to reach maximum biomass occur during summer.

In addition to seasonal variations, sponges were affected by severe episodic events. The series of strong storms that struck central California from November 2022 to April 2023 coincided with a far more dramatic decrease in sponge cover than the previous year in both intensity and duration. While the decline of sponge cover began after mid-October in both years, the decline in the winter of 2022–2023 was far faster than the previous (2021–2022) winter and led to the complete disappearance of above-ground Hymeniacidon perlevis at APN and KP up through the end of sampling, July 2023. Sponge populations at RBR did not completely disappear and some above-ground biomass remained before growth resumed in May 2023. One possible explanation is that the storms covered such a long period that while sponges were able to close their ostia to survive increased turbidity and/or low salinity temporarily (Parker 1910), they were unable to feed or survive prolonged exposure. However, the only significant statistical correlation with precipitation was at KP, where sponge cover lagged 1 month behind. Other parameters, most notably temperature, have been implicated in driving H. perlevis’ phenology due to correlation with cover (Stone 1970; Gaino et al. 2010; Cao et al. 2012), but the specific mechanism triggering sponge regression remains an open question for further study. Considering the number of regions H. perlevis now inhabits (Samaai et al. 2022), understanding these reactions to severe weather is important in the context of ongoing climate change.

While a seasonal life cycle is common in populations of Hymeniacidon perlevis around the world, there can be exceptions. H. perlevis grew rapidly and continuously with no seasonal pattern in areas near to a nutrient-rich wastewater outflow from a fish farm while sponge populations further from the inlet showed greater variability (Mercurio et al. 2023). These observations support a plastic seasonality that is variable and may not occur under stable conditions. Within Elkhorn Slough, the population dynamics of H. perlevis across years and between sites demonstrates the importance of microhabitats, perhaps related to food availability and water flow as noted by Mercurio et al. (2023). These relationships are crucial to predicting the abundance and timing of H. perlevis in all habitats, artificial and natural.

Given seasonal changes in sponge cover, the impacts of Hymeniacidon perlevis on Elkhorn Slough are unlikely to be uniform over the year. Pumping and biomass may not scale as tightly with cover as we have shown here in our simple model (Fig. 4, Suppl. material 2: table S3), since sponges may vary in shape or in proportion of biomass that is buried in the mud. However, in the absence of more detailed measurements, it seems reasonable to assume that pumping rates and biomass increase with cover. Thus, the environmental impacts of H. perlevis, on water filtration and biomass, are likely greatest during mid-October in this estuary, when cover is highest, and smallest to nonexistent during March to May, when cover is lowest.

The impacts of filtration by Hymeniacidon perlevis on Elkhorn Slough water are not known. In laboratory experiments, H. perlevis consumed bacteria, reducing concentrations by 38–90% in laboratory conditions (Longo et al. 2010; Maldonado et al. 2010; Zhang et al. 2010; Longo et al. 2022), making it a potent feeder on bacterioplankton. Elkhorn Slough is characterized by eutrophication and nutrient runoff from agriculture throughout the watershed that lowers water quality (Caffrey et al. 2002a, Caffrey et al. 2002b, Hughes et al. 2011). At the densities H. perlevis achieves, especially at peak biomass in the fall, this grazing capability may affect bacterial densities in Elkhorn Slough. If H. perlevis is in a site with abundant bacteria as food, it may persist and affect water conditions year-round. Feeding rates vary with water temperature, with peak removal rates at 15 °C and less removal at higher and lower temperatures (Zhang et al. 2010). In those same mesocosm experiments, elevated water temperatures also resulted in greater ammonia released into the water by the sponges, suggesting H. perlevis may play a role in both carbon and nitrogen cycling within the Elkhorn Slough and other coastal habitats.

Hymeniacidon perlevis may provide novel habitat as it grows, or in contrast be a competitor for space. This species can grow rapidly in a period of months, which along with its resilience likely enables it to outcompete native species for space. Large aggregations of sponges can alter a boundary layer that provides unique overlying habitat (Gili and Coma 1998). The sponges may also serve as novel habitat: amphipods and polychaetes were seen within the tissues. Both temperate and tropical sponges have long been recognized as habitat for a variety of fish and invertebrate species (Wulff 2006), with relationships ranging from mutualistic (Siemann and Turco 2023) to predatory/parasitic (Magnino and Gaino 1998). However, many of these endobionts are specialists that seek out a particular host species (Rützler 2012) as opposed to the presumably opportunistic invertebrates making use of an introduced sponge in Elkhorn Slough. Recent experimental evidence suggests that brittle stars, polychaete worms, crabs, and fish all act as spongivore predators on H. perlevis in Argentinian Patagonia where it has also been introduced (Buch et al. 2024). Species that might consume H. perlevis in Elkhorn Slough are unknown, though observations from British Columbia suggest that the native nudibranch Doris montereyensis Cooper, 1863 grazes on H. perlevis (Harbo RM, personal communication). While the sponges, as novel, seasonally available habitat and prey may shelter and/or provide food for invertebrate and fish species, the improvised nature of these relationships suggests they may be subject to rapid change. H. perlevis may provide distinct food and habitat compared to other Elkhorn Slough sponges, including the introduced Cliona celata Grant, 1826, Halichondria (Halichondria) bowerbanki Burton, 1930, and Chalinula loosanoffi (Hartman, 1958), the cryptogenic Haliclona (Reneira) cinerea (Grant, 1826), and unidentified Topsentia sp. Berg, 1899, as well as the native Halisarca sacra de Laubenfels, 1930 and Mycale (Mycale) macginitiei de Laubenfels, 1930 (Wasson et al. 2001; Caffrey et al. 2002b).

The spectrum of impacts described above complicate an assessment of whether Hymeniacidon perlevis should be considered invasive, or if its seasonal and interannual dynamics make its impacts too variable over time. Estimating the seasonal abundance of H. perlevis in a specific site can be a useful first step in adaptive management for this species. If removal or reduction of H. perlevis populations is required, then timing culling efforts to periods when sponges are visually present but not yet reproductive (i.e. June-August for population dynamics such as observed at Elkhorn Slough) may provide the most efficient removal method. This is not possible once the species is widely established, such as at Elkhorn Slough, but could be possible soon after detection in a new area. Additionally, understanding seasonal cycles may facilitate efforts to model the spread of H. perlevis to new regions. Such season-informed models have been generated using phenology data of non-native species before, such as the brown marmorated stink bug, Halyomorpha halys (Stål, 1855) (Kamiyama et al. 2021; Reznik et al. 2022) and the Asian shore crab Hemigrapsus sanguineus (De Haan, 1835) (Giménez et al. 2020). These phenological models are particularly important for predicting how non-native species may alter their ecosystems in the face of climate change (Colautti et al. 2017) or anomalous climate events, such as the severe storms that affected H. perlevis’ phenology during this study.