Latitudinal investigation of Olympia oyster elevational distribution

The goal of our investigation was to characterize the upper elevational limit of Olympia oysters (Ostrea lurida) across the full latitudinal range of the species, and to determine how this limit varies as a function of air temperature. At each of 26 sites from British Columbia to Baja California, we located the highest live Olympia oysters to identify the upper distributional limit at that site. Where Pacific oysters (Magallana gigas) were present, we also identified their upper elevational limit. In addition, we identified the upper limit of acorn barnacles at the site as a biological indicator of the upper limit of sessile invertebrate communities in the intertidal zone, analogous to the treeline on mountains. We examined how the upper elevational limits of the two oyster species and barnacles varied across latitudes and as a function of air temperature, obtained from a nearby weather station. Given that other studies have shown negative effects of extreme heat events on intertidal organisms, we hypothesized that the elevational distribution of these species would be compressed at sites experiencing more frequent high temperatures. Thus, we expected the upper limit of these species to be lower at hotter, lower latitude sites. However, timing of the tides affects exposure to heat events and sites where the lowest tides in summer occur near midday are exposed to more heat stress than sites where they fall at night or early morning. Consequently, we expected that temperatures during low tide exposure would be a better predictor of the upper limits than temperatures across all time periods. To further explore whether high temperatures might cause mortality near the species’ upper limit, we also compared the elevations of the highest dead vs. live Olympia oysters and examined size distributions in relation to elevation for this species. For example, if a recent thermally induced die-off had occurred, dead oysters should occur higher than live oysters, and small, newly settled oysters should occur above larger, older ones. We also expected that the highest oysters at a site would be in shaded rather than sun-exposed microhabitats. Ultimately, the goals of this investigation across an expansive latitudinal range were to improve understanding of the ecology and climate sensitivity of Olympia oysters to inform future restoration planning for this important foundation species, and to serve as a model for evaluating climate resilience of other foundation species that occur along steep environmental gradients.

Methods
Geographic scope

We conducted surveys at 26 sites located in 18 estuaries across almost the entire biogeographic range of Olympia oysters, more than 21 degrees latitude, from 30.5° N to 51.6° N. Sampling occurred May-November 2023, mostly in summer. This effort was a collaborative project with more than 100 people involved in the field. We chose sites that had appropriate substrate for oysters distributed across a range of elevations, at least from Mean Lower Low Water (MLLW) to Mean Higher High Water (MHHW); we thus could be confident that the absence of oysters from high elevations was not due to lack of suitable substrata.

Characterizing intertidal distributions

Olympia oyster elevational patterns

Our primary focus at all 26 sites was to identify the upper elevational limit of Olympia oysters –  to identify their landward boundary in the intertidal zone. At all sites, we placed markers on the highest 10-20 live Olympia oysters that could be found. We only included oysters that were in fairly exposed locations; that is, we excluded oysters that inhabited deep crevices, under cobble, or in seeps or pools that retain water at low tide. These oysters were at least 1 m apart from each other within 100 m of shoreline. In these relatively rapid field surveys for oysters, we would not have detected newly settled juveniles <0.05 cm in size.  
We measured the elevation of the highest Olympia oysters flagged at each site. At most sites we used a networked RTK GPS. At the most northern two sites and one Oregon site we used laser leveling from the waterline and estimated relative tidal elevations using the predicted waterline elevation at that time. At the three most southern sites, we used differential GPS. We surveyed a landmark (known/stable ground control point) at the start and end of our fieldwork at the site, and in most cases these measurements differed by <3 cm.

To compare the upper limits of oysters across estuaries with different tidal ranges and different vertical datums (CGVD2013 converted to Chart Datum in Canada, NAVD88 in US and WGS84 in Mexico), it was critical to use a consistent universal currency. For each site, we obtained an estimate of MLLW and MHHW in the same vertical datum as the elevation measurements. In the US the estimates were provided using NOAA’s vDatum (https://vdatum.noaa.gov/). The only exception within the US was at San Dieguito Lagoon where the nearest NOAA water level station was not representative of local water level conditions; in this case, MLLW and MHHW were calculated using water level data collected with HOBO water level loggers deployed by the San Onofre Nuclear Generating Station Mitigation Monitoring Program during 2021. In Canada, MLLW and MHHW were calculated from the closest government tide station (https://www.tides.gc.ca/en/stations). In Mexico, MLLW and MHHW were calculated using the astronomical tidal models from CICESE (https://predmar.cicese.mx/calendarios/ ), and the local tidal datums were validated against benchmarks defined by the Mexican Navy (Secretaría de Marina). 

We used the following formula to standardize all elevation measurements to a relative scale that takes into account tidal range:

Relative elevation of measurement 

= (field elevation measurement - elevation of MLLW ) / 

(elevation of MHHW - elevation of MLLW)

Critically, all four elements in this formula were in the same vertical datum for a site, e.g. Chart Datum in Canada but NAVD88 in USA. On this scale, MLLW has a relative elevation of 0, MHHW of 1.

Because we were interested in the upper tolerance limit at each site, we used the highest 5 elevation values for each site that were within 0.1 unit on the relative scale of the highest measurement.  We preferred 5 measurements to a single one, to avoid anomalous outliers. However, at sites with very low oyster densities, sometimes there were fewer than 5 oysters near the upper limit, so in these cases some of the top 5 measures spanned a large elevational range. To obtain a more accurate estimate of the true upper limit, we only used those within 0.1 elevational unit.

At those sites where the tide was low enough after conducting the above surveys, we also characterized the elevational zone with the densest Olympia oysters, which was always lower than the upper limit. At 12 sites we placed quadrats (15 x 15 cm) in 10 of the densest oyster locations at the site and counted the oysters. We took the elevation of these 10 quadrats as described above.

Pacific oyster and barnacle upper limits compared to Olympia oysters

At a selection of sites where time and tide permitted, we measured the elevation of the 10 highest live Pacific oysters using the same methods described above. At most (22 of 26) sites, we measured the elevation of the 10 highest barnacles, which represented the highest sessile invertebrates present at the site. At sites with both oyster species present, individuals <7 cm in size were sacrificed following measurement to view diagnostic features for definitive identification. Individuals >7 cm were assumed to be Pacific oysters.

Relationships among air temperature, latitude, and Olympia oyster upper limits

We used the nearest reliable weather stations to obtain hourly air temperature for the five years prior to and including our field surveys (2019-2023) for each site. In a few cases, there were gaps in the datasets. Where these gaps lasted a week or more (South Sequim Bay and Brickyard Park, San Francisco Bay), we substituted data from the next closest weather station. We were unable to fill a 3-month gap (Jan-Mar 2023) at one site (RV Park, Estero Punta Banda).

To examine air temperatures during low tide exposure, we also filtered the above dataset of all air temperatures to only include hours when tide level was below MLLW. We used the nearest tide stations for water level estimates to obtain predicted hourly tide levels relative to MLLW for each site for the same five-year period. 

We were interested in the role of extreme temperatures in shaping the intertidal distribution of oysters, and in exploring which methods of calculating extreme temperatures would be most ecologically relevant. We therefore examined multiple metrics for characterizing temperatures that were extreme across the range of the Olympia oyster. We recognize that these “extreme” temperatures were not extreme on a global scale, but rather only in relative terms of what organisms at these sites experience. To characterize exposure to these air temperature extremes, we calculated the number of temperature values <5°C and >30°C both before and after filtering data for water levels below MLLW. Due to occasional missing data, we divided that by the number of hourly readings present to get the proportions of readings below 5°C and above 30°C. Totals and proportions were for the entire 5-year period of 2019-2023. We used 5°C and 30°C as thresholds for extreme temperatures based on inspection of temperatures across the range; we had initially considered using 0°C and 40°C, but determined they occur so rarely at most sites as not to be useful for correlative analyses. As an alternative approach, we also examined the 10th and 90th percentile of temperatures for each site both before and after filtering data for water levels below MLLW.

Within-site variation in upper elevational limits

We collected and analyzed ancillary data to determine whether within-site variation in Olympia oyster elevations showed patterns that would be consistent with mortality associated with extreme high temperature events.

Sun exposure

If high temperatures set the maximum elevation of an intertidal species, upper distributional limits should be related to exposure to sunshine. For the highest Olympia oysters at each site, we assigned each oyster to one of three levels of exposure, 1) shady: oyster would be mostly shaded at midday, for instance under overhang; 2) exposed: oyster would be mostly in full sun at midday, for instance on top of rock; and 3) partial: somewhere in between shaded and exposed. At some sites, there was no variation (e.g., all of the highest oysters were exposed); at other sites, low tides during our field survey occurred at night and exposure could not be assessed.

Evidence of recent die-off

If recent die-offs caused by low tide exposure to heat occurred, one would expect to find dead oysters above live oysters at a site. In addition to searching for the highest live Olympia oysters, field teams also searched for the highest dead Olympia oysters when time permitted. We focused on recently dead individuals that were gaping (oysters still cemented to the substrate by the bottom valve, with the top valve still attached). As with live oysters, we identified individuals at least 1 m apart within 100 m of shoreline, and measured elevations as described above.

If the local populations experienced periodic die-offs of the highest oysters due to heat waves, the highest oysters at a site should be smaller than lower ones. At all except one site, we measured shell sizes (longest dimension) for the highest Olympia oysters while collecting their elevations.

Substrate size

Large substrates heat up more slowly than small ones during low tide exposure, so if high temperatures were setting the upper limit, the highest oysters might be limited to the largest substrates at the site. We measured the size (maximum linear dimension) of the substrate to which the highest live Olympia oysters were attached. Unattached single oysters were assigned an arbitrary substrate size of 1 cm and oysters on bedrock were assigned a substrate size of 1000 cm.