From 2010 to 2020, we surveyed M. gigas densities and size frequency distributions at 4 sites each in Newport Bay and San Diego Bay, California, USA (Fig 1, Table 1). Sites were surveyed two to six times throughout the decade (median = 3). The habitats among the sites varied and included pier piles, seawalls, riprap, cobble, mudflat, and chain link fence (Table 1). A California Department of Fish and Wildlife scientific collecting permit (SCP-3027) required to complete this research was issued to author D. Zacherl and all appropriate approvals were obtained for the research.

All surveys occurred within the intertidal zone during low tide at tidal elevations ranging from -0.49 m to +1.81 m MLLW, although the tidal range surveyed varied across years and sites, with broader ranges of elevations surveyed in more recent years, particularly in San Diego Bay. Standardization of tidal ranges for statistical comparisons among years is addressed below (see Statistical Analyses, Table 1); data recorded at tidal elevations not included in this study are available upon request to the corresponding author.

From 2010–2016, we surveyed replicate randomly placed 0.5 m x 0.5 m (0.25 m²) quadrats along a 50-m transect placed in the middle of the visible oyster zone parallel to the water line. We surveyed randomly selected elevations for each quadrat placed within a 1–4 m band above and below the transect line; the band width depended upon the slope of substrate surveyed. Tidal elevation was estimated for the entire transect at each location with RTK-GPS or a LaserMark LM-30, with maximum and minimum elevations recorded (see Table 1 for tidal range surveyed). For pier piles, we did not lay a transect, instead we randomized the pile and side of pile surveyed, and elevation surveyed.

Once we established that tidal elevation played a significant role in oyster distributions [41], we recorded tidal elevation of every randomized quadrat in all subsequent surveys (2017–2020). We also modified sampling on habitats with steep slopes (seawall, pier pile and fence habitats) to more precisely account for tidal elevation by using quadrats sized 0.5 m length X 0.15 m height (0.075 m²). These smaller-area quadrats are as long as those used prior to 2017, but not as high, thus sampling a more constrained tidal range on a vertical surface. In 2020, we sampled across a much broader tidal elevation range relative to previous years (e.g., see Table 1 for tidal ranges surveyed) but constrained among-year comparisons of density data to commonly surveyed tidal elevations.

Within each quadrat, we identified and counted all live M. gigas that were at least 50% inside of the quadrat. We used internal and external shell characteristics to identify M. gigas. External characteristics that differentiated NIS M. gigas from native O. lurida (the only other oyster species present in our surveyed estuaries) included larger maximum height, thicker body, and the presence of shell foliations. Internal characteristics included a chalky white interior shell color and absence of chomata; presence of chomata is a generic-level trait for Ostrea [42, 43]; also see Fig 1 of [24]. We recorded the heights and lengths of as many M. gigas encountered as was practical given low-tide time constraints (n = 31–308 oyster lengths recorded per survey, median = 112); height of each oyster was measured from the umbo to the furthest lip of the oyster shell. When oyster densities were low (i.e., prior to 2015), every oyster encountered was measured. When oyster densities were high (2015 and later) it became impossible to measure all oyster heights within a single tide; surveyors were instructed to systematically work across the quadrat and only measure the heights of the first 5 oysters encountered.

Oyster shell heights were used to visualize changes in size frequency distributions before and after 2015 as an indicator for recruitment through time. Shell height measurements were recorded at three sites in Newport Bay (FS, LI, and AC), and one site in San Diego Bay (CV). Heights were binned into 10 mm increments (i.e., 10–19 mm, 20–29 mm, etc.) and distributions were compared within each site before and after 2015. Shell heights < 10 mm were excluded. Oyster sizes from recruitment plates in San Diego Bay, deployed yearly from May through September [44] indicate that oysters less than ~40 mm can generally be considered Year 0 (young of the year) and those > 40 mm considered Year 1 or older, though growth is known to vary broadly across sites [45] based upon food availability, temperature and salinity [46].

We downloaded publicly available sea surface temperature (SST) data from the Shore Stations Program [47] at Newport Beach Pier (open coast near Newport Bay, 33.606414°N, 117.929997°W), Scripps Pier (open coast near San Diego Bay, 32.866692°N, 117.257496°W), and from the National Data Buoy Center [48] for NOAA station IIWC1 at the US Midway (inside of San Diego Bay, 32.713889°N 117.176667°W). All SST temperature data spanned from January 1, 2010, to December 31, 2020. Some SST data were missing from each year at Newport Beach and Scripps Piers, especially during June (18 days) and July (11 days), 2014 at Newport Beach Pier. Temperature data at Newport Beach Pier were recorded once per day sporadically between 06:00–18:00 (Pacific time), at Scripps Pier once per day between 07:00–18:00 (Pacific time), and hourly at NOAA station IIWC1. Since M. gigas is known to settle and recruit in southern California during summer from June to September [44], we focused our temperature data exploration during those months. We reviewed the daily SST data from 2010 through 2020 to look for a noticeable difference in summer water temperatures that may have correlated with increasing local M. gigas densities.

To account for discrepancies in tidal elevation sampled among survey years, with some earlier survey years constrained to a smaller subset of elevations surveyed, we standardized the ranges of elevations so that consistent elevations were compared within each site across years, but not necessarily across sites (e.g., elevations surveyed at CV were lower than those at GC, see Table 1 for the full range of sites surveyed versus the standardized range used in statistical analyses). For each site, we examined whether oyster density changed across time using 1-way ANOVA (time treated as categorical factor) in JMP v16.0.0. We did not combine sites into the same statistical analysis because each site had a unique history of years surveyed and standardized elevations that could be used. Most data were log-transformed [Log (X+ 1)] to address heteroscedasticity prior to analysis (Table 2). The log-transformations greatly improved issues with heteroscedasticity and non-normality as was evidenced by visual inspection of plots of residuals by predicted, studentized residuals, and residual by normal quantiles. Statistical analyses were not performed on the SST and size frequency distribution data; qualitative trends were elucidated via visual inspection of the data.

From 2010 to 2020, M. gigas density increased by 2.2 to 32.8 times at 7 of the 8 sites surveyed across two southern California bays (Fig 2).

In Newport Bay, M. gigas densities significantly increased over time at all 4 sites. At AC, density increased 2.7 times from 4.53 ± 1.64 (SE) to 12.27 ± 2.48 oysters/m² (Tables 1–3, Fig 2, ANOVA, p = 0.0033). At CI, density increased 23.2 times from 0.53 ± 1.64 to 12.27 ± 2.48 oysters/m² (Tables 1–3, Fig 2, p = 0.0112). At LI, density increased 2.2 times from 11.45 ± 1.26 to 25.28 + 3.21 oysters/m² (Tables 1–3, Fig 2, p < 0.0001). At FS, density increased 32.8 times from 1.75 ± 0.63 to 57.32 ± 8.38 oysters/m² (Tables 1–3, Fig 2, p < 0.0001).

In San Diego Bay, M. gigas densities increased over time at 3 of the 4 sites. At PN, density increased from 0.00 ± 0.00 to 8.35 ± 4.55 oysters/m² (Tables 1–3, Fig 2, p = 0.0127). At CV, density increased 4.1 times from 7.00 ± 3.08 to 28.33 ± 4.36 oysters/m² (Tables 1–3, Fig 2, p = 0.0026). At GC, density was stable over time and averaged from 35.20 ± 12.44 to 44.13 ± 8.74 (1 SE) oysters/m² (Tables 1 and 2, Fig 2, p = 0.0994). At GB, density increased 3.8 times from 58.09 ± 9.69 to 222.67 ± 34.03 oysters/m² (Tables 1–3, Fig 2, p = 0.0003).

From 2010 through 2015, M. gigas in the 10-mm size class, representing very recent Year 0 recruitment, were not detected at the three sites surveyed in Newport Bay nor at CV in San Diego Bay (Fig 3). All sites showed increased proportions of this size class following 2015, ranging between 0.8% at FS and 4.2% at AC. At three of four sites (LI, AC and CV), there were also higher proportions of oysters (6%– 9.3%) in size classes from 20 mm to 40 mm (also likely Year 0) relative to before 2015. Unlike at the other three sites, proportions of oysters in most of the smaller size classes (20–50 mm) at FS in Newport Bay decreased between 0.6% and 19.9% after 2015. All sites showed evidence of Year 0 size classes (< 40 mm, see methods) both before and after 2015 (Fig 3).

Overall, median summer SST increased between 2010 and 2014 and has been consistently warmer since 2014 (Fig 4). In both Newport Bay and San Diego Bay, the coolest average monthly temperatures occurred in 2010. From 2010 to 2014 there was a clear 2–5°C shift upwards in the median monthly temperatures, most notable during the summer months of July–September (Fig 4). The greatest shifts in median water temperature occurred in August and September, with 4°C shifts in Newport Bay, and 3–5°C shifts in San Diego Bay. Between 2010–2014, Newport Bay and San Diego Bay saw an increase in the average SST (not shown) during summer months (June–Sept) of 3–4°C and 2–3°C, respectively. After 2014, median temperatures in all summer months varied among years, but were consistently warmer by ~2–3°C relative to the 2010–2014 time period.