Evidence of exceptional oyster‐reef resilience to fluctuations in sea level

Abstract Ecosystems at the land–sea interface are vulnerable to rising sea level. Intertidal habitats must maintain their surface elevations with respect to sea level to persist via vertical growth or landward retreat, but projected rates of sea‐level rise may exceed the accretion rates of many biogenic habitats. While considerable attention is focused on climate change over centennial timescales, relative sea level also fluctuates dramatically (10–30 cm) over month‐to‐year timescales due to interacting oceanic and atmospheric processes. To assess the response of oyster‐reef (Crassostrea virginica) growth to interannual variations in mean sea level (MSL) and improve long‐term forecasts of reef response to rising seas, we monitored the morphology of constructed and natural intertidal reefs over 5 years using terrestrial lidar. Timing of reef scans created distinct periods of high and low relative water level for decade‐old reefs (n = 3) constructed in 1997 and 2000, young reefs (n = 11) constructed in 2011 and one natural reef (approximately 100 years old). Changes in surface elevation were related to MSL trends. Decade‐old reefs achieved 2 cm/year growth, which occurred along higher elevations when MSL increased. Young reefs experienced peak growth (6.7 cm/year) at a lower elevation that coincided with a drop in MSL. The natural reef exhibited considerable loss during the low MSL of the first time step but grew substantially during higher MSL through the second time step, with growth peaking (4.3 cm/year) at MSL, reoccupying the elevations previously lost. Oyster reefs appear to be in dynamic equilibrium with short‐term (month‐to‐year) fluctuations in sea level, evidencing notable resilience to future changes to sea level that surpasses other coastal biogenic habitat types. These growth patterns support the presence of a previously defined optimal growth zone that shifts correspondingly with changes in MSL, which can help guide oyster‐reef conservation and restoration.


| INTRODUCTION
Climate change poses a significant threat to ecosystems across the globe with pronounced impacts to biogeography, manifesting most prominently at the edges of species ranges, near an organism's threshold tolerance to physicochemical or biotic controls. Changes to the environment along these boundaries could result in a variety of outcomes including species adaptations (Hoffmann & Sgro, 2011), changes to phenology (Edwards & Richardson, 2004;Poloczanska et al., 2013), range shifts (Chen, Hill, Ohlemüller, Roy, & Thomas, 2011;Davis & Shaw, 2001;Poloczanska et al., 2013), community and trophic restructuring (Edwards & Richardson, 2004;Walther et al., 2002), and even localized extinction (Colwell, Brehm, Cardelus, Gilman, & Longino, 2008;Pinsky, Worm, Fogarty, Sarmiento, & Levin, 2013). The magnitude of these responses will depend on an organism's sensitivity to the suite of environmental factors that may be undergoing change or the resultant altered biotic relationships, the rate at which the system is changing (Ackerly et al., 2010), and the reaction time of the species to adapt.
The response and reaction time of various organisms to climate fluctuations are highly specific among different taxa. Conditions detrimental to fitness may occur if there is a notable lag in community response to climate alterations, as seen with forest communities and temperature (Bertrand et al., 2011), or a species unable to shift correspondingly to the vector and acceleration at which an environmental variable, such as temperature or average rainfall, is changing (Burrows et al., 2014;Dobrowski et al., 2013;Zhu, Woodall, & Clark, 2012). The nature of environmental shifts across geographic space means mobile organisms can respond more readily by migrating (Pinsky et al., 2013), whereas sessile organisms must rely on adaptation, propagation, and habitat modification to maintain their populations (Bertrand et al., 2011). As many communities depend on the persistence of habitatforming foundation species, it is crucial that these sessile ecosystem engineers keep pace with climate changes to sustain habitat area and quality (Colwell et al., 2008;Kirwan & Megonigal, 2013;Ridge et al., 2015).
Biogenic habitats are experiencing environmental change in a variety of forms, including temperature, precipitation/desertification, ocean acidification, and sea-level rise (SLR); all of which vary in rate geographically and can interact to cause complex responses in ecological communities as populations react differently (Tingley, Koo, Moritz, Rush, & Beissinger, 2012). While many of these climatic factors shift laterally across a geographic space, SLR also presents change in the vertical, which is particularly important for developed coastal areas where infrastructure prevents upland migration. Intertidal and shallow subtidal biogenic habitats exist in a narrow elevation range due to a combination of biophysical intolerance and interspecific interactions (Bertness & Ellison, 1987;Fodrie et al., 2014;Paine, 1971). Fluctuations in sea level can represent a dramatic change to species that are relegated to intertidal zones, such as saltmarshes and mangroves, because it changes the inundation time during a tidal cycle. The change in sea level may be significant compared to the overall range of elevations the organisms occupy. If these foundation species cannot maintain their surface elevations compared to relative SLR (RSLR, the combination of eustatic sea-level rise and local shifts to continental crust), they will become imperiled by the stress of saltwater submergence, which could result in a loss of their supported communities and associated ecosystem services (Kirwan & Megonigal, 2013;Lovelock et al., 2015).
While sea level along the coast of the United States is generally rising 2-6 mm/year, it fluctuates significantly from year to year, seasonally, and even on shorter timescales (weeks to months). These changes can range from 15 to 20 cm interannually with the most dramatic being greater than 30 cm (Morris et al., 2002;Sweet, Zervas, & Gill, 2009;Sea Level Trends, NOAA Tides & Currents). Some of this variation is due to seasonal temperature and wind climate, but pronounced deviations may also arise with the complex interconnectivity of the North Atlantic Oscillations (NAO), prolonged or frequent storm activity, and sea-level anomalies linked to the strength of the Gulf Stream (Ezer, 2016;Ezer, Atkinson, Corlett, & Blanco, 2013;Goddard et al., 2015;Kolker & Hameed, 2007;Sweet, Zervas, & Gill, 2009). Losada et al. (2013) demonstrated that interannual shifts in sea level in other areas of the Atlantic Ocean can be on the order of 4-12 cm, with ENSOinduced sea-level shifts exceeding historical RSLR and an increased frequency in sea-level extremes occurring in recent decades. Shortterm elevations in sea level are responsible for more frequent flooding along the U.S. East Coast (Ezer & Atkinson, 2014) and increased coastal erosion (Theuerkauf, Rodriguez, Fegley, & Luettich, 2014). These fluctuations in sea level may have a marked impact on coastal and estuarine habitats as their regularity and longevity are expected to increase (Ezer & Atkinson, 2014).
The persistence of biogenic habitats, along with the critical services they provide to ecosystems and coastal infrastructure, is uncertain in the face of accelerated RSLR. Vegetated habitats (saltmarshes, mangroves, and seagrasses) alter their surface elevations through passive trapping of sediment from the water column, accumulation of annual aboveground biomass, and by augmenting belowground biomass forcing the sediment surface upwards (Morris et al., 2002). Habitats constructed by invertebrates (e.g., coral reefs, oyster reefs, worm reefs) rely on individual growth and gregarious settlement to maintain their placement in suitable conditions, with multiple generations building on one another. Given the range of accretion rates exhibited by these ecosystem engineers (Baustian, Mendelssohn, & Hester, 2012;Bhomia, Inglett, & Reddy, 2015;Cahoon et al.,2006;Perry et al., 2015;Sasmito, Murdiyarso, Friess, & Kurnianto, 2016), many will maintain their relative position with moderate rates of RSLR, while higher rates of RSLR may result in massive loss of coastal habitats along large geographic stretches due to drowning and compression against coastal infrastructure (Pontee, 2013).
Oyster reefs are ubiquitous features within temperate and subtropical estuaries, spanning from the intertidal to subtidal zones depending on salinity and climate (Baggett et al., 2015;Walles et al., 2016). While oysters provide many important benefits to the ecosystem, populations are recovering from decimation during the last century (Beck et al., 2011). Previous work examining intertidal oyster-reef growth indicates that constructed Crassostrea virginica reefs have a relatively high growth capacity compared to other coastal habitats , far outpacing any predicted rate of RSLR. However, growth rates are highly variable across reefelevation gradients due to stress associated with exposure (desiccation) and submergence (competition and predation), with the reef crest and base exhibiting stunted or lack of growth (critical exposure boundaries) and the sides growing at the highest rate (optimal growth zone [OGZ], Ridge et al., 2015) (Figure 1). In the lower portions of estuaries, where salinities are typically greater than 30 ppt, C. virginica reefs cannot persist in the subtidal zone due to overwhelming predation and competition by species that are intolerant to exposure Powers et al., 2009), indicating that transitioning from intertidal to subtidal conditions will place reefs that cannot keep pace with rising seas in peril (Ridge et al., 2015).
While oyster-reef growth patterns are well constrained over decadal scales, their sensitivity to changing sea levels over monthly to yearly timeframes is still relatively unknown. Considering the degree to which sea level can fluctuate from weeks to months, understanding how the critical boundaries and OGZ will shift in response is necessary information for proper timing and siting of oyster restoration projects as well as assessing how future trends of RSLR will affect reef persistence, which in many estuaries is the only available hard substrate.

| Study area
This study was conducted using C. virginica oyster reefs located in the

| Study design
Our study included an assortment of reefs of different ages (grouped into three "generations") ranging from 2 years to a century old ( Figure 2b). Constructed reefs ranged from 5 to 10 m in diameter, similar in size to many natural reefs in our study area, while the natural reef included in this study was one of the larger reefs locally, approximately 15 × 50 m (width × length). Constructed reefs began as mounds of loose, recycled oyster shell (cultch shell) measuring 3 × 5 × 0.15 m (width × length × height), followed by natural recruitment of oyster larvae from the estuary. Growth (cm/year) of three reefs constructed over a decade prior (Grabowski, Hughes, Kimbro, & Dolan, 2005; hereafter "decade-old reefs"), including one reef con- To examine fine-scale growth across oyster reefs, terrestrial lidar (Riegl LMSZ210ii laser scanner) was used to image reefs ( Figure 4). Terrestrial laser scanning followed previously reported methods , using RTK-GPS-positioned reflectors to georectify the point cloud to less than 1-cm horizontal and 1.5-cm vertical accuracy. Elevations were recorded in the North American Vertical Datum of 1988 (NAVD88). Reef-mapping with lidar required dry weather and a low spring tide, providing only a narrow operating window to scan a reef, which typically took an hour. As such, it sometimes required several days to several months to acquire all the scans of each reef generation, particularly in the case of the young reefs, which were numerous and widely separated (denoted by the bar widths in Figure 3a). Within each reef generation, we collected scans during the same season and normalized the data to annual rates to avoid uneven seasonal influences across the time steps. The combination of RiSCAN Pro (Riegl) and F I G U R E 1 Reef growth conceptual model adapted from Ridge et al. (2015) that predicts oyster-reef growth rate with aerial (tidal) exposure. Relevant elevations in NAVD88 are provided for aerial exposures (%) for the Cape Lookout region of North Carolina. The lower critical exposure boundary occurs where oyster-reef growth equals the rate of relative SLR (RSLR), shifting correspondingly as RSLR changes. Oyster-reef growth is illustrated (right panel) across a hypothetical reef-elevation profile using dotted (time 1) and solid (time 2) profile lines data were transformed (cube transformation) to meet assumptions for parametric analysis, and a series of t-tests were run to compare monthly mean water temperatures and salinities between scan periods within each reef generation.

| Water level and quality
Monthly MSL data from 2009 to 2015 indicate that sea level in the study area was −0.028 ± 0.062 m NAVD88 (mean ± SD) ( Figure 3b).
Prior to the start of scanning, the study area experienced prolonged levels of high water from frequent sea-level anomalies during the fall and winter of 2009-2010 that persisted for 5 months .  (Tables 2 and S1).

| DISCUSSION
Changes in reef morphology and reef-wide growth were tightly aligned with month-to-year patterns in sea level (Figures 5-7, Table 2). The magnitude and direction of these interannual fluctua- Variations in water quality did not appear to have a strong influence on reef growth, but fluctuations in salinity may be responsible for nonconformities in the expected growth pattern. Water temperatures over the entire study period did not display dramatic deviations that would explain differences in growth between years (Figure 3c).
Overall temperatures varied about a degree Celsius or less between scan periods, having cooler temperatures during the second time steps. Cooler temperatures would be associated with less growth in C. virginica (Dame, 1972), which is contrary to our results, indicating that temperatures had little effect on how growth manifested on reefs.
In contrast, salinity decreased throughout the entire study period, which could have impacted how growth manifested at lower elevations along the reef profiles. Crassostrea virginica is robust to fluctuations in salinity, and the range (15-36 ppt) is not outside of the Eastern oyster's tolerance (Shumway, 1996). Salinities below 25 ppt may actually be more conducive for oyster growth, especially in subtidal waters (Walles et al., 2016) where fresher water may hinder predators (e.g., gastropods) and competitors (e.g., macroalgae), which could explain high growth below the OGZ on the decade-old reefs after periods of pronounced lower salinity. Both salinity and seasonal temperature cycles can influence oyster reproduction and recruitment, and this can result in interannual variability in larval settlement patterns (Ortega & Sutherland, 1992). While we did not collect oyster spatfall data for this study, other research conducted within our study area and adjacent waters has shown larval availability to be high within the estuaries of North Carolina and that interannual variability of larval supply is relatively low in the more saline sounds (Carroll, Riddle, Woods, & Finelli, 2015;Ortega & Sutherland, 1992;Puckett & Eggleston, 2012).
In these waters, the major determinants of oyster recruitment are the T A B L E 2 Summary of peak growth, sea level, temperature, and salinity for each time step by reef generation postsettlement processes of predation and competition (Carroll et al., 2015;Fodrie et al., 2014).
Our study was not designed as a controlled experiment to control for a suite of abiotic and biotic factors. However, changes to biotic or abiotic influences, other than salinity, should generally impact the magnitude of growth profiles, rather than shifting growth curves upward or downward as we observed. For instance, increased thermal stress or disease should decrease the magnitude of a growth curve overall but not change the elevations associated with growth. The exceptions to this response would be those processes dictated by tidal-exposure stressors (e.g., desiccation, predation, and competition), which shift correspondingly with sea level, reinforcing sea level as the primary control (and certainly the most parsimonious based on the available data). Erosion was most likely the response of these reefs returning to equilibrium after a year of high water preceding the first scan (Figure 3b).

| Decade-old reefs
This would have temporarily increased the growth ceiling before waters returned to a lower stand, exposing the reef crest to higher desiccation stress and potentially greater foraging by avian predators (American Oystercatcher, Haematopus palliates), resulting in oyster mortality. Thus, it appears that while oysters cement together to create a solid reef matrix, oyster mortality within the taphonomically active zone (layer of living oysters) due to overexposure could compromise the outer reef structure, making it more susceptible to erosional forces. A similar process has been documented on coral reefs during extreme low tides (Anthony & Kerswell, 2007), and it vertically mirrors dieback of marshes in response to long term over inundation creating highly reduced soils (Koch, Mendelssohn, & Mckee, 1990). Even though oysters cement together, death of an oyster can lead to the valves disconnecting as the adductor muscle releases, which can result in loose shell if wave energy eventually works the valves apart. When larger oysters die within the taphonomically active zone, bigger clumps of oysters can be worked loose and displaced if multiple smaller oysters are attached to a large valve. These areas then appear as erosion spots when the reef is mapped again if the space has not been reoccupied by other oysters in that time. While present, loose oyster shell on or around the reef was not quantified nor is it likely to have a high residence time on the reef crest as these areas are subject to the greatest hydrodynamic energy, and shells are often observed scattered across the adjacent sandflat.
Therefore, it is difficult to use loose shell as a metric for erosion on these isolated reefs. We could also have witnessed a natural process of compaction within the reef. As multiple generations of oysters continue to build on one another, the structure of the reef matrix likely condenses and fills the empty cavities of once living oysters. During years of average or higher water, this process is likely compensated by oyster growth on the reef surface and therefore only manifests as loss during periods of protracted low water. Some places, such as areas of the Chesapeake Bay, experience shell loss through dissolution in more acidic conditions; however, our study area is not experiencing pronounced acidification, and dissolution is an unlikely source of reef surface erosion.

| Young reefs
Elevations of growth maximums on young reefs also paralleled changes in sea levels ( Figure 6). Sea levels during 2012 were higher than the initial time step for the decade-old reefs, which resulted in an elevated OGZ. When sea levels dropped 2 cm in 2013, maximum growth also occurred 2 cm lower, doubling the average accretion rate while also increasing growth along deeper areas of these reefs. Young reefs appear to have the strongest response to sea-level changes, but this could be a result of our scan-period resolution isolating narrow time frames with fairly distinct trends in sea level. Areas of the young reefs exhibited growth as high as 8-11 cm/year after their construction in 2011 . The sustained high average growth across these reefs indicates they will only require 4-6 years to occupy the accommodation space and reach MSL. Volume changes on young reefs are an order of magnitude greater than the decadeold reefs because of the greater surface area located within the OGZ coupled with a much higher vertical accretion rate. This pattern of growth, two to three times greater than the decade-old reefs, follows the modeled maturation of other intertidal habitats such as marshes, which experience rapid growth during immaturity that asymptotes at the rate of RSLR at maturity (Allen, 1990;Jennings, Carter, & Orford, 1995).

| Centennial reef
Growth changes on the centennial reef over the study period behaved comparably to patterns displayed by the other study reefs (Figure 7).
Peak growth during the initial time step occurred at the base of the Lower growth peak occurs in OGZ predicted OGZ. Similar to the decade-old reefs, the centennial reef experienced erosion at elevations above −0.07 during a period of relatively low water (2013). However, the higher water of 2014-2015 yielded 4 cm/year accretion in previously eroded areas at or above MSL, which manifests as a comparably large increase in volume across the reef plateau. This growth rate (4 cm/year) is equivalent to the 3.9 cm jump in average sea levels over the two time intervals and is also comparable to the rapid growths displayed by the young reefs. Thus, mature reefs not only follow the intertidal oyster-reef growth paradigm, they also have the capacity to respond just as rapidly to changes in sea level as immature reefs.
This would indicate that mature oyster reefs are not confined to nearly asymptotic growth at the rate of RSLR like that of other coastal habitats.
Considering the clarity in response of the centennial reef to the relatively Growth on the centennial reef was measured at higher elevations than the other reefs included in this study. This reef is much larger than the constructed reefs, and we may be witnessing a certain degree of facilitation (Bruno, Stachowicz, & Bertness, 2003) within the oyster population, similar to the Northern Acorn Barnacles of the New England rocky intertidal (Bertness, 1989), due to thermal buffering and reduced desiccation stress. In fact, each time the centennial reef was sampled there remained ponds of water on the reef's plateau during low tide (appear as dark spots on the reef in Figure 2b). Presence of these ponds shows that the reef is fairly nonporous, retaining water at higher elevations throughout a tidal cycle. This would indicate that, while the OGZ magnitude of large natural reefs corresponds to shifts in sea level, the elevations at which the OGZ manifests may behave differently as a reef matures and expands.
As a larger reef spread across a greater exposure gradient, the degree to which the centennial reef would have been subject to forces impacting surface elevation may have been different than the constructed reefs.
Having a much larger mass, the reef may be more prone to compaction in years of low water. Its size and proximity to sandy shoreline may have also fostered higher foraging by the American Oystercatcher during years when the reef was more exposed, although this is likely not as significant as desiccation mortality and reef compaction. Alternatively, the centennial reef is in a more sheltered environment and encompasses a much larger population of oysters compared to the constructed reefs and is adjacent to other reefs of similar size. This would provide a greater localized source of oyster larvae, helping maintain recruitment even during years of lower larval supply relative to the smaller, more exposed constructed reefs (O'Beirn, Heffernan, & Walker, 1995, 1996.

| Resilience to sea-level fluctuations
Oyster reefs appear to be in dynamic equilibrium with sea level. Like other intertidal habitats, decadal or longer measurements of mature reef surface elevation changes would show they track RSLR (DeAlteris, 1988). However, unlike other habitats, annual rates of change in reef vertical relief could be ±5 cm depending on relative sea levels.
Prolonged shifts in sea level cause different reef elevations to essentially turn off or on, akin to a phenomenon present in coral reefs (Perry & Smithers, 2011) but operating at a much greater magnitude. Rapid coral reef vertical accretion is on the order of 0.5-0.9 cm/year during reef turn-on (Perry & Smithers, 2011), while oyster reefs can achieve greater than 2 cm/year regardless of maturity. Therefore, oyster reefs, despite being sessile organisms, are well adapted for tracking this particular climate velocity vector as long as the environment remains estuarine. It should be noted that this outstanding vertical accretion has only been measured on intertidal oyster reefs, and subtidal oyster reef accretion may not respond to fluctuations in sea level. Existing in the intertidal zone mediates the impact of biotic interactions and near bottom hypoxia, to which subtidal reefs are exposed (Lenihan, 1999), allowing for growth to be dictated primarily by sea level and aerial exposure regime assuming no disease or degraded water quality.
The ubiquity of this response across oyster-reef ages is a testament to their resilience to RSLR as well as their utility and longevity for stabilizing shorelines, likely reducing the potential impacts of the coastal squeeze (Pontee, 2013). This work supports the value of using the OGZ for intertidal oyster population management (Ridge et al., 2015), being an effective predictive tool for oyster-reef growth patterns. Use of the OGZ will prove highly valuable in restoration projects, particularly the implementation of green infrastructure such as living shorelines that incorporate oyster breakwaters. However, there remains a need to measure oyster-reef lateral expansion and adjacent benthic-sediment modification processes. These measurements will help establish whether or not oyster reefs will build landward as they track SLR or eventually create reef islands. Similar to wetland upland transgression with RSLR (Kirwan, Walters, Reay, & Carr, 2016), this process will depend on the ability of oysters to expand up the littoral slope . Considering the outstanding vertical growth captured in this study, the primary limiting factor would appear to be the rate of transgression (i.e., expansion up slope) at a particular shoreline.
This study presents evidence that intertidal oyster reefs are highly responsive to short-term fluctuations in local sea level even at maturation. When compared to other coastal habitats and their capacities for RSLR response, oyster reefs are unparalleled in their ability to maintain surface elevation with changing sea level. Greatest recorded rates of surface elevation change in intertidal and shallow subtidal systems such as marshes, mangroves, and corals are below 1-2 cm/year excluding storm-related allochthonous sedimentation (Baustian et al., 2012;Bhomia et al., 2015;Perry et al., 2015;Sasmito et al., 2016).
Overall, this research further solidifies that oyster reefs are resilient habitats that will become increasingly important in estuarine systems with changing sea level.

ACKNOWLEDGMENTS
We would like to thank the members of the Coastal Geology Lab and