Near-daily reconstruction of tropical intertidal limpet life-history using secondary-ion mass spectrometry

Ecology of Yellowfoot limpet

In the Tropical Pacific, sympatric limpets (Cellana melanostoma, Cellana exarata, Cellana sandwicensis, Cellana talcosa) inhabit the Hawaiian rocky intertidal ecosystem, where they graze on crustose coralline algae (CCA) and epibenthic microorganisms. Distribution ranges from the splash zone (upper-intertidal) to subtidal zone, and across the entire Hawaiian Archipelago²⁶. They are dispersed across the majority of seamounts, atolls, and islands, however, not all species are present in every rocky intertidal locality, which reflects species-specific micro-habitat preferences.

The reproduction cycles for each species appears to vary in time and space, and on-going long-term monitoring efforts are in progress to define this critical life-history trait. Previous studies on the yellowfoot limpet C. sandwicensis, reveal that reproduction is highly synchronized from December to March^27,29. Gametogenesis also occurs from June to August, however, the level synchronicity and intensity of this second spawn period are inconsistent.

These limpets are gonochoristic and considered to be sequential hermaphrodites⁴⁴. The sex ratio is near 1:1(M:F) during spawning season, however, we have directly observed populations to maintain disproportionate sex ratios.

Development of this broad-cast spawning limpet has been described from egg to post-larvae, where settlement occurs in less than 4 days post-fertilization²⁹. This short larval duration ensures recruitment to the same localized intertidal environment, and reduces likelihood of hybridization between sympatric species with similar life-histories²⁶.

For wild limpets, growth rates shift through ontogeny—average monthly growth decreasing from 4–5 mm shell length (SL) as juveniles to 2–3 mm SL as adults²⁷. Limpets also exhibit seasonal growth patterns—influenced by temperature and feeding^28,37. Currently, growth rates of large individuals (>50 mm SL) and species longevity are absent in the literature.

Regional climate and coastal oceanography

Ka’alawai is located on the south-facing shoreline of Oahu Island, Hawai’i (21°15’20.7“N 157°47’30.8“W). This area, defined as a rocky intertidal zone, is primarily comprised of basalt outcrops, boulders and benches, and supports a diverse community of epibenthic flora and fauna. The area is relatively easy to access by foot, and has been continuously exposed to various anthropogenic factors, which includes development, urban run-off, and subsistence fishing.

The microclimate of the region is characterized by mild, wet winters (January to March) and dry, hot summers (July to September). The mean daily atmospheric temperature range and mean daily sea-surface temperature range are 18.44–31.38 °C and 22.67–30.18 °C, respectively. The annual precipitation is low relative to windward sides of the island, with maximum rainfall of 6.35 cm (data sources: US climate station USC00519397: Waikiki 717.2; PacIOOS Nearshore Sensor 04 (NS04): Waikiki Aquarium). Although freshwater input from precipitation along this coastline is considered to be marginal, the mixing of submarine groundwater discharge generates a unique geochemical profile for surface seawater at Ka’alawai. In particular, the mean surface salinity for this study site has been reported to be 25.4 ‰, which reflects this highly localized land-sea interaction⁴⁵.

The coastal oceanography of this region is predominantly influenced by wave, wind, and tidal forces. The south-shore region experiences a mixed tidal cycle—having both diurnal and semi-diurnal sinusoidal constituents per lunar day—with a tidal range of 58 cm and 91 cm during neap tide and spring tide, respectively; The trade winds from north-easterly direction (between 22.5°–67.5°) account for ~63% of the year with mean annual intensity around 5 m/s;⁴⁶ and South swells with wave amplitudes of ~3 m are generated by storms in the Tasmanian Sea during Northern Hemisphere Summer^47,48.

Modern and historical specimens

On June 28th of 2018, live Yellowfoot limpet (Cellana sandwicensis) specimens CW1 and CW2 were collected from the rocky intertidal zone at Ka’alawai, Oahu, Hawai’i (Fig. 7). The animals were immediately sacrificed/dissected using scalpel blade, and measured for shell dimensions using a caliper. Limpets were weighed to determine gonadosomatic index, and gonads were preserved for histological examination. Shells were rinsed in an ultrasonic bath and air-dried.

A historical specimen BPBM (identification number 250851-200492) was loaned from the Bernice Pauahi Bishop Museum Malacology Department Collection. This specimen’s geographical and ecological origin is unknown, but was identified as C. sandwicensis by its characteristic shell morphology⁴⁹. This specimen was selected for its large size to estimate life-expectancy of this limpet species, as well as to evaluate this method for paleoclimatology studies.

Permission was not required to obtain specimens used in this study, and limpets were collected at a size exceeds the legal minimum shell length of 31.8 mm (Hawaii State Law is enforced by Department of Land and Natural Resources). Ethical approval was not required to conduct analysis.

Characterization of shell microstructure

Shell microstructure was identified before isotopic analysis could be attempted. Each shell was cross-sectioned from anterior to posterior direction using a low speed saw (Isomet 1000, Buehler) equipped with a 0.5 mm diamond coated blade. Parallel cuts were made at the apex or maximal growth-axis to obtain two replicate 1.3 mm thick-sections per specimen. The first replicate thick-sections, prepared for micro-sampling, were further cut into <15 mm long pieces and mounted on a single glass round, and the second replicate thick-sections, prepared for sclerochronology, were mounted in its entirety on a large glass slide. Specimens were mounted in using quick-drying epoxy (EPO-TEK 301, Epoxy Technology Inc, Billerica, MA) set in a mold, grinded with F1000 grit SiC powdersecosecondar, and polished with 3 and 1 µm Al₂O₃ powder on a lapping wheel. Polished sections on glass rounds were then sonicated, rinsed with methanol, and carbon coated to ~250 Å (Cressington Carbon Coater 208carbon, Watford UK). Microstructures of unstained specimens were identified by Scanning Electron Microscopy (SEM; JEOL JSM-5900LV, USA) photomicrograph following MacClintock⁵⁰.

Raman spectroscopy was used to characterize biogenic carbonate mineralogy by comparing shifts in relative peak position and intensity between calcite and aragonite polymorphs⁵¹ (see Supplementary Fig. 4). A silicon wafer standard was used to determine spectral center when grating was 800 grooves/nm. Single spectrum analysis was performed in each microstructure layer using a green laser at 514 nm. A total of six (n = 6) sampling sites were selected haphazardly for a given microstructure layer, which comprised of ten accumulations averaged across 10,000 s. The shell’s exterior surface layer did not return clear spectral peaks, and thus has been excluded from our analysis.

Secondary ion mass spectrometry analysis

Hawaiian limpet Cellana sandwicensis shells CW1, CW2, and BPBM were analyzed for oxygen isotopes. Polished thick-sections were imaged by light and scanning electron microscopy (SEM) to guide sampling by ion microprobe. We sampled in the crossed-foliated, calcite layer (M + 2) to avoid mixing calcite and aragonite layers, and sequential measurements were performed moving from the shell margin toward the apex along the growth axis.

To achieve sub-weekly resolution for an annual δ¹⁸O cycle, modern specimens, CW1 and CW2, were analyzed using an average interval of 252 µm and 288 µm between samples, respectively. To achieve sub-annual resolution across multiple δ¹⁸O cycles, the historical specimen, BPBM 250851-200492, was analyzed using an average interval of 554 µm. The total number of microprobe sample measurements was 55, 43, and 56 for CW1, CW2, and BPBM, respectively. A single, annual isotope cycle was analyzed for the modern specimens, and four annual isotope cycles were analyzed for the historical specimen.

The carbon-coated thick-sections were placed under vacuum conditions to prevent contamination prior to being measured by CAMECA-IMS-1280 ion microprobe (SIMS; W.M. Keck Research Laboratory, University of Hawai’i) for oxygen isotopes. The primary ion beam Cs+ was set at 2.5 nA for 120 s presputtering. Ions were extracted at ~8 kV. Microprobe rastered across a 15 µm² area, which accounted for 1-3 lunar daily growth increments. Each measurement included 30 cycles with 10 s integration period. The ¹⁶O and ¹⁸O were measured in multicollection mode using two Faraday cups with 10¹⁰ and 10¹¹ Ω registers, respectively. The b-field was controlled by nuclear magnetic resonance. Mass resolving power was 1958 % min^-1, which allows detection of possible interference ions. To correct for instrumental isotope mass fractionation, University of Wisconsin Calcite, UWC-3 (δ¹⁸O = 12.49‰ Vienna Standard Mean Ocean Water – VSMOW), was measured consecutively before and after performing microprobe analyses for each specimen (n = 12)⁵². Corrections were made based on these groups of standard measurements with limited monitoring for instrumental drift. The reproducibility measurements (2σ) of UWC-3 reference material ranged from 0.17 to 0.35‰, which reflected measurement precision and reproducibility of standard measurements for same-day measurements. Measurement errors are reported as 2σ, which reflects both precision (2 standard error) and reproducibility (2 standard deviation) (see Supplementary Data 1 and Supplementary Data 2). Following the microprobe analyses, shell samples were imaged under SEM to expose sample scars for sclerochronology.

Predicted shell δ¹⁸O

To examine if Hawaiian limpet shells are in isotopic equilibrium with their environment, we compared measured δ¹⁸O_calcite to predicted δ¹⁸O_seawater calculated from seawater-surface temperatures (SST) and surface seawater salinity (SSS). Sea-surface temperatures were obtained from in-situ PacIOOS Nearshore Sensor 04 (NS04): Waikiki Aquarium, Oahu, Hawai’i at 21.26587° N, -157.82275°W (Sea-Bird Electronics model 37SMP, ±0.002 accuracy from -5 to 35 °C). Predicted values were calculated using the equilibrium fractionation equation for calcite and water described by Friedman and O’Neil:⁵³

Where T is temperature in Kelvin and α is the fractionation between calcite and water from the equation:

Where δ¹⁸O_calcite and δ¹⁸O_seawater are in VSMOW. Conversion from VPDB to VSMOW scales were performed using the following equation:⁵⁴

The relationship between mean annual surface seawater δ¹⁸O and mean annual surface salinity for the Tropical Pacific region (50-year data set) was used to calculate δ¹⁸O_seawater (Legrande & Schmidt 2006):

Mean surface seawater salinity was 25.4 ‰ for Ka’alawai⁴⁵.

The temperature error for proxy measurements for CW1, CW2, and BPBM were±1.54,±1.53, and±1.60 °C, respectively.

Temporal alignment of shell δ¹⁸O

Following SEM and SIMS, we removed the carbon coat with methanol and Kimwipe tissue. The same thick-sections analyzed by microprobe were stained following the previously mentioned procedure, and imaged by light microscopy. The SEM and light photomicrographs were over-layed using Photoshop to accurately expose the spatial-relationship between SIMS-analyzed points and growth lines/increments. For temporal alignment of isotope measurements, we used predicted δ¹⁸O_calcite values that aligned with major lines as anchoring points. Calendar dates (based on a lunar year) were assigned for every isotope measurement between these anchors by counting micro lines (lunar daily growth). We applied previous research on growth and reproduction to resolve alignment discrepancies between predicted δ¹⁸O_calcite and measured δ¹⁸O_calcite.

Climate reconstruction of historical shell

The exact location from which the historical specimen BMBP was collected from is unknown. Climate was reconstructed from the shell isotope record—assuming that δ¹⁸O_calcite is precipitated in equilibrium with δ¹⁸O_seawater. Sea surface temperature was calculated from sequentially sampled isotope measurements across ecologically relevant salinity values. Based on the profile with the most biologically relevant temperature thresholds (min and max), we predicted historical sea surface salinity (SSS).

Growth measurements

The polished over-sized, thick-sections were stained with Mutvei’s solution to expose major lines, micro lines, and micro increments by light microscopy^55,56,57. Shell thick-sections were placed in a petri dish and submerged in Mutvei’s solution for 45 min held constant at 37–40 °C with constant stirring. These stained thick-sections were imaged using Nikon Eclipse E600 Polarizing light microscope at 100x magnification for performing growth line measurements.

Daily growth was measured along two axes using the standard measuring tool in ImageJ. The first type of daily growth was measured between two micro increments along the growth axis. The second type of daily growth was measured along the horizontal axis (anterior to posterior orientation).

For the latter, we recorded x-coordinates for each point where a micro increment band intersected the M + 3 layer, and subtracted x-coordinates of sequential points to calculate horizontal distance or growth. Back-calculated shell length measurements were used to model age-at-length data (see Supplementary Eq. 2, Supplementary Fig. 2).

We analyzed sub-monthly growth of modern specimens across an annual isotope cycle to understand temporal changes in growth. This period was selected for based on the shell length at which C. sandwicensis enters adulthood, which allows interpretation of adult growth (Kay et al. 1983).

Shell growth model

We used the back-calculated measurements of shell length-at-age as data inputs to estimate parameters of the von Bertalanffy growth function (VBGF), a standard method of describing growth in marine animals^58,59. The VBGF (Eq. 5) was fit to shell length (in mm) at age (in months) data for each shell individually (i.e., CW1, CW2, BH) and pooled samples using non-linear parameter estimation:

where L(t) is length (mm) at age t (years), L_∞ is the mean asymptotic length (mm), K is the growth coefficient, and t₀ is the theoretical age at length zero. Growth curves were fit by constraining ({L}_{0},)to a common shell length of settlement in order to increase the accuracy of VBGF parameter estimates⁶⁰. The length at post-larval settlement (i.e.,(,{L}_{0})) of Cellana sandwicensis was obtained from existing literature as 0.254 mm²⁹ and equated with ({t}_{0}) = -0.09026. To determine if measurements from all shells could be pooled for a single growth model, we used likelihood ratio tests to test for pairwise differences in ({L}_{{{{{{rm{infty }}}}}}}) and K between shells by generating a χ² statistic for each set of comparisons (sensu;^61,62 Haddon, 2011). Growth data for the pairwise comparisons were truncated to a shell length range of 0 – 45 mm that represented the range of data overlap for all three shells to minimize bias from the larger maximum size of shell BPBM⁶¹. We used R v3.3.1 (R Core Team, Vienna, Austria), and Excel v2013 (Microsoft Corporation, Redmond, WA, USA) to perform the growth model statistical analyses.

Statistical analysis

Unless stated otherwise, all statistical analyses were accomplished in SAS (SAS v9.2, SAS Institute Inc., Cary, NC, USA). Pearson’s correlation coefficients were computed using Proc Corr to determine linear relationships between SST_calculated and SST_measured. We also used correlation analysis to describe monthly growth by changes in growth frequency and DG_SL, respectively. To analyze daily growth as a function of time, we performed univariate ANOVA with repeated measures using Mixed Proc. Pair-wise comparisons of unequal group sizes was performed using Tukeys-Kramer post-hoc analysis. Significant differences were determined using an alpha of 0.05 for all statistical procedures.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Source: Ecology - nature.com