Ocean warming is the key filter for successful colonization of the migrant octocoral Melithaea erythraea (Ehrenberg, 1834) in the Eastern Mediterranean Sea

Study site, temperature, coral collection and sclerite isolation

Specimens for SEM examination, crystallography, and carbon and oxygen stable isotope analysis were collected from both the Mediterranean Sea and the Red Sea (Fig. 1). Melithaea erythraea branches were collected during scuba dives in June 2016 and May 2017 in Nahsholim Bay. Samples from Hadera port were collected in June 2016, and from Eilat (Red Sea), branches were collected in May 2017. Branches from M. biserialis were also collected from Eilat in May 2017. From all collected colonies, a branch from the distal parts of the colony was removed using scissors and preserved in absolute ethanol (96%) prior to examination. For comparison, a specimen of M. erythraea from Fine et al. (2005), collected in Hadera port, was examined as well (collected by Y. Aluma in 2002 and stored in formalin at the Steinhardt Museum of Natural History, Israel National Center for Biodiversity Studies, Tel Aviv, Israel).

Branches from each test colony were sub-sampled (1 cm in length). Sclerites were separated from the soft tissue by placing each sub-sample in Eppendorf tubes filled with 10% sodium hypochlorite until the soft tissue was dissolved. After 30 min, the organic debris was removed, and the sclerites were rinsed with distilled water several times to wash off the excess bleach and supernatant. Multiple sclerites were ground to homogenous powders for later isotopic analysis.

Information of ambient conditions (temperature) was collected from Israel Oceanographic and Limnological Research (IOLR) monitoring station in Hadera. For the period of 1994–2004, temperature measurements were taken from a bottom-mounted Paroscientific-8DP060 ADCP. For the period of 2004–2018, temperature measurements were taken from a bottom-mounted 600 kHz WorkHorse Monitor ADCP. The ADCP was located at 11.6 m until 2004, then relocated to 26 m depth, southwest of the easternmost edge of the coal terminals in Hadera. The reported temperature sensor precision is ± 0.4 °C. The data series was curated for outliers, smoothed and binned.

All samples were collected under the permit from the Israel Nature and Parks Authority (permit number: 2016-18/42200). Conducting the surveys in their areas was approved by Port of Hadera Authority and Eilat-Ashqelon Pipe-Line Company.

Isotope ratio mass-spectrometry (IRMS)

The fractionation of the oxygen isotopes into biogenic carbonates is a function of ambient temperature and isotopic composition of the seawater at the time of their formation (Grossman & Ku, 1986; McConnaughey, 1989a; McConnaughey, 1989b; Kim & O’Neil, 1997). If the isotopic ratio of ¹⁶O and ¹⁸O (expressed as δ¹⁸O) of both calcium carbonate and water is known, the temperature at deposition can be calculated. The isotopic fractionation of carbon (δ¹³C) in skeletal material provides information about the organism’s metabolism, as well as nutritional information (McConnaughey et al., 1997) and, thus, can provide insights into the calcific response of migrating calcifying species to severe oligotrophic conditions.

Stable isotope (δ¹⁸O and δ¹³C) measurements on bulk powders were performed at the stable isotope laboratory of the Max-Planck Institute for Chemistry, Mainz, on a Thermo Delta V mass spectrometer interfaced with a Gasbench preparation unit. Sample digestion took place on-line, in >99% orthophosphoric acid, at 70 °C. Coral samples were analyzed together with several calcite standards, including the international standard IAEA603. The reproducibility of these routinely analyzed in-house CaCO₃ standards is typically ≤0.1‰(1 SD) for both carbon and oxygen isotope ratios. Both δ¹⁸O and δ¹³C of the sclerites are reported relative to the Vienna Pee Dee Belemnite (VPDB) standard scale.

Estimation of CaCO₃ depositional temperatures

The oxygen isotope composition of biogenic CaCO₃ is a function of ambient water temperature and the δ¹⁸O of the seawater at the time of its formation and from this depositional temperatures can be estimated (Grossman & Ku, 1986; Kim & O’Neil, 1997). The calcite temperature-dependent fractionation during bio-mineralization is described by the equation of Friedman & O’Neil (1977): (1)${\displaystyle 10^{3}ln{\alpha }_{calcite-water}=2.78\left(10^6∕T^2\right)-2.89}$ where α_{calcite−water} is the oxygen isotope fractionation factor between calcite and water (Eq. (2)) and T is the water temperature (K). (2)${\displaystyle {\alpha }_{calcite-water}\mathrm{ }=\left(10^3+{\delta }^{18}{O}_{calcite}\right)∕\left(10^3+{\delta }^{18}{O}_{water}\right)}$

The mean annual δ¹⁸O_sw value of 1.6‰  (Sisma-Ventura, Yam & Shemesh, 2014; Sisma-Ventura et al., 2016) and 1.9‰ (Mizrachi et al., 2010) of surface water in the Mediterranean and the Gulf of Aqaba, respectively, and the δ¹⁸O_SC of M. erythraea from both habitats were used for the calculation of deposition temperatures. It is noted that the δ¹⁸O_sw in both the Mediterranean Sea and the Gulf of Aqaba fluctuated by less than 0.5‰, annually. A combination of the analytical uncertainties of >0.5‰ for both measurements translates into uncertainty of ∼2 °C (i.e., 0.2‰ °C ⁻¹).

Estimation of percentage of metabolic carbon intake during calcification

We estimated the percentage of the metabolic carbon that contributed to δ¹³C_SC using the mass balance equation (McConnaughey et al., 1997): (3)${\displaystyle {\delta }^{13}{C}_{calcite}-{\epsilon }_{calcite-bicarbonate}=M\left({\delta }^{13}{C}_{Food}\right)+\left(1-M\right){\delta }^{13}{C}_{DIC}}$ where M is the percentage of the metabolic carbon contribution and ε_{calcite−bicarbonate} is the enrichment factor between calcite and bicarbonate (+1‰, Romanek, Grossman & Morse, 1992). The δ¹³C of M. erythraea sclerites may indicate a food source (phytoplankton-small zooplankton (Zeevi-Ben Yosef & Benayahu, 1999); δ¹³C ≈ −20‰), and DIC (1‰; Mizrachi et al., 2010; Sisma-Ventura, Yam & Shemesh, 2014; Sisma-Ventura et al., 2016) of both habitats were used to calculate the metabolic contribution to the skeletal buildup.

X-ray diffraction crystallography

Crystallinity is an important parameter of mineral aggregates such as skeletons. This property can be effected by internal heterogeneity in the crystal, nucleation rate, protein framework structure, Sr and Mg concentration and crystal growth rate and is, therefore, a useful parameter to understand environmental effects expressed by in the calcification.

X-ray diffractometry (XRD) was used to evaluate the crystallinity of the sclerites. Full width at high maximum (FWHM) of crystallinity level (Patterson, 1939) was calculated for the calcite’s d₁₀₄ peak following the Scherrer equation (Scherrer, 1918): (4)${\displaystyle \tau =\frac{K\lambda }{\beta cos\theta }}$ where τ is the mean size of the ordered (crystalline) domains, λ X-ray wavelength; K is the shape factor; β is FWHM and θ is the Bragg angle. As all parameters are constant other than the FWHM, then τ is proportional to β⁻¹. Given that the shape factor could not be determined in most case, FWHM can be used as an index to the level of crystallinity.

The analysis was conducted with a Rigaku MiniFlex benchtop XRD, with the sclerites deposited from suspension on a custom slide and allowed to dry in a desiccator. Diffraction was carried out from 10 to 75° at 0.01° steps at a rate of 2.15° per minute.

Sclerite morphology and statistical analysis

The morphometric complexity of the sclerites was assessed by SEM analysis. Spindle-shaped sclerites from all specimens were placed in non-coated high/low vacuum mode and were examined and photographed with Jeol JCM-7000 NeoScope benchtop SEM with secondary electron and backscatter modes set to magnify at ×200 and ×300. From each specimen, n = 32–46 sclerites (see Table 1) were examined, and the measurements of the axes (length and width) were taken using built-in software. In addition, along the longitudinal axis, warts were counted to obtain their density. The number of warts was divided by axis length (in µm).

Statistical analyses were performed under permutation concepts, non-parametric tests that analyze quantified data that do not satisfy the assumptions underlying traditional parametric tests (e.g., normality, etc., Collingridge, 2013). To compare the effect of “site” (Nahsholim, Hadera, and Eilat), and “axis” (long, short) on sclerite morphometrics, we performed a nested Permutation ANOVA, followed by a pairwise permutation posthoc test. Before analysis, data were normalized. Axes’ ratio and wart density on sclerites among sites were examined with one-way Permutation ANOVA, followed by pairwise permutation posthoc test. All values are presented at a confidence interval of 95%. All statistical and multivariate analyses were performed with R i386 3.3.3 (R Core Team, 2014) using ‘lmPerm’ and ‘rcompanion’ packages.

Sclerite morphometrics

The sclerites (shown in Fig. 2) captured in the SEM were mostly spindle shaped, a shape found in all samples. In the Red Sea, we also observed a spheroidal shape (for both M. erythraea and M. biserialis). The sclerites exhibited some visual differences between the sites. The Red Sea sclerites looked thicker than the Mediterranean ones, and their wart density appear to be higher (see Table 1, raw measurement of the sclerites is provided in the Supplemental Information).

The sclerite axis characteristics were significantly different by size (Fig. 3A, Nested permutation ANOVA, p < 0.001), where posthoc pairwise comparisons suggested that the corals in the Red Sea have thicker sclerites in comparison with all other test corals collected in the Mediterranean (all pairings with the Eilat specimen, except Eilat-Hadera 2002, are p-adjusted < 0.002).

These differences are also reflected in the long-to-short axes ratio comparison (Fig. 3B, Permutation ANOVA, p < 0.001), where the Red Sea sclerites had the smallest ratio (3.19 ± 1.15) compared to all other sites (4.78 ± 1.18, and 4.98 ± 1.52 in Nahsholim 2016 and 2017, respectively, and 4.55 ± 1.30, 4.14 ± 1.30 in Hadera 2002 and 2016, p-adjusted < 0.005), not including the M. biserialis pair (p-adjusted = 0.39). M. biserialis’ ratio (3.58 ± 2.23) was similar to both Hadera (2002 and 2016) specimens (p-adjusted > 0.2) and was significantly different than all the other sites (p-adjusted < 0.02). In addition, the Hadera 2016 sclerite axis ratio is significantly different than those collected from Nahsholim (p-adjusted < 0.05).

Warts along the sclerites’ axis were counted to determine their density. The wart density differed among the different sites (Fig. 3C, Permutation ANOVA, p < 0.001), where the M. erythraea Red Sea specimen has a significantly higher density (0.057 ± 0.013 # µm⁻¹) than those collected from the Mediterranean (0.049 ± 0.006, 0.051 ± 0.007, 0.0047 ±  0.008 in Nahsholim 2016–2017, and 0.047 ± 0.006 and 0.049  ± 0.006 # µm⁻¹ in Hadera 2002 and 2016, respectively, p-adjusted < 0.05), except the pair collected from Eilat and Nahsholim 2017 (p-adjusted > 0.1).

Isotopes

The results in Fig. 4 summarize the δ¹⁸O_SC and δ¹³C values of M. erythraea and M.  biserialis sclerites from the Gulf of Aqaba (Red Sea), and of M. erythraea from the Israeli coast (SEMS). The δ¹⁸O_SC value of samples from the Gulf of Aqaba, which were collected during May 2017, ranged between 0.33 ± 0.045‰  in M. erythraea and 0.01 ± 0.061‰  in M. biserialis (the average temperature of deposition was 23 °C and 24.5 ± 0.5 °C, respectively). Similarly, M. erythraea samples from the Port of Hadera (the first documented colonization), was collected in the spring of 2002, and yielded δ¹⁸O_SCvalues between 0.26 and 0.4 ± 0.063‰ (depositional temperatures of 21 ± 0.7 °C). The δ¹⁸O_SC values of samples collected during 2016 and 2017 in early spring from Hadera and Nahsholim Bay ranged between 0.22 ± 0.03 and 0.36 ± 0.052‰(deposition temperatures of 21.2 ± 0.5 °C). The 2017 (early summer) samples from the Nahsholim Bay yielded δ¹⁸O_SC values of 1.07 ± 0.05‰  (deposition temperatures of 27.7 ± 0.4 °C).

The samples from the Gulf of Aqaba (May 2017) yielded δ¹³C_SC values ranging between 0.25 ±  0.033‰  (M. biserialis) and −0.26 ±  0.02‰  (M. erythraea), and those sampled from the Hadera port in 2002 averaged −0.38 ± 0.1‰. The δ¹³C_SC values of samples collected during 2016 and 2017 in early spring from Hadera and Nahsholim Bay ranged between −1.38 ± 0.062 and −1.16 ± 0.12‰. The 2017 early summer samples from the Nahsholim Bay yielded δ¹³C_SC values, averaging −1.24 ± 0.18‰. Results are summarized in Table 1 and the ‘Isotopes’ section within the Supplemental Information.

Crystallography

Full width at high maximum (FWMH) of the d₁₀₄ peak of the calcites ranged from 0.10 to 0.27 with the d spacing ranging from 2.988 to 2.997; peak asymmetry ranged from 0.56 to 2.8. The Eilat samples and 2002 Hadera samples exhibit the lower values of asymmetry and d spacing values with the higher FWHM values relative to the 2016 and 2017 values. The length of δ¹³C_SC and the long axis are positively correlated to FWMH (inversed to crystallinity, Table 1, Fig. 5).

Sea surface temperature (SST)

Mean SST in the coast of Hadera had not changed significantly since the early 1990’s and remained at 23.2 ± 4.3 °C (n = 199229). However, this figure is misleading as the extreme temperatures have shifted in both summer and winter (Fig. 6). Notably, minima temperature (10th percentile) has increased by ∼2 °C while maxima temperature (90th percentile) has diminished by ∼1 °C. In addition, the fractional time during which temperature was below 18 °C (1994’s 20th percentile) has diminished from ∼18% in 1994 to less than 5% in 2019, while the fraction above 28 °C (1994’s 80th percentile) has diminished from ∼24% in 1994 to ∼15% in 2019.