Quantifying the Effect of Anthropogenic Climate Change on Calcifying Plankton

Widely regarded as an imminent threat to our oceans, ocean acidification has been documented in all oceanic basins. Projected changes in seawater chemistry will have catastrophic biotic effects due to ocean acidification hindering biogenic carbonate production, which will in turn lead to substantial changes in marine ecosystems. However, previous attempts to quantify the effect of acidification on planktonic calcifying organisms has relied on laboratory based studies with substantial methodological limitations. This has been overcome by comparing historic plankton tows from the seminal HMS Challenger Expedition (1872–1876) with the recent Tara Oceans expedition material (2009–2016). Nano CT-scans of selected equatorial Pacific Ocean planktonic foraminifera, have revealed that all modern specimens had up to 76% thinner shells than their historic counterparts. The “Challenger Revisited” project highlights the potential of historic ocean collections as a tool to investigate ocean acidification since the early Industrial Revolution. Further analyses of such biotic archives will enable researchers to quantify the effects of anthropogenic climate change across the globe.

carbon cycling in the ocean 14 . They are not only responsible for a quarter or more of global carbonate production 15 , but also an integral component of the marine food chain 16 . However, to date there is little information available with regards the impact of OA on calcification rates of planktonic foraminifera. While lab-based studies are invaluable in assessing short-term controls on calcification in certain species, they do not allow for adaptation in planktonic foraminifera, which have to date, never reproduced in a laboratory environment 16 .
Attempts to evaluate the impact of OA on calcification in the oceans have been hindered by an inability to directly compare the calcification capability of today's plankton species with equivalent specimens from exclusively pre-and early-industrial times 17 . Field-based studies have tried to compare the calcification ability of a modern plankton species today with their pre-industrial counterparts by using seafloor sediments to represent 'pre-industrial' times. However, one centimeter depth of deep sea sediment may represent 100's of years of Earth's history 18 , therefore this method inevitably incorporates a mix of specimens from a large window of geological time, leading to potential circularity when comparing data from these sediments to modern sediment traps or plankton tows.
Multiple plankton tows collected by the HMS Challenger crew from across the world's ocean basins between 1872-1876 1 provides a resource that resolves issues highlighted in both lab-based and field-based studies; the tow material contains the planktonic foraminifera, of a known age, that were alive in the open ocean at the time of sampling. Thus, the HMS Challenger collection provides exceptional baseline data for 19th-21st century ocean acidification evaluation. Here we present the results of a focused study from the central Pacific Ocean (Fig. 1), a region identified as being vulnerable to ocean warming, ocean acidification, and ocean deoxygenation with increasing atmospheric CO 2 levels 19,20 . The results of this study highlight the utility of museum collections for studies of anthropogenic climate change, and demonstrates an urgent need for more empirical data.

Results
Here we test the utility of these collections for modern climate change studies by comparing specimens of planktonic foraminifera collected in tow nets from HMS Challenger Station 272 (1875) (Fig. 1) with Tara expedition (2011) 21 stations 127 and 128, in the eastern Pacific Ocean; a region that is recognized as being particularly vulnerable to deoxygenation and surface warming in the wake of anthropogenic climate change 20 . Both Tara and HMS Challenger stations were sampled during September thus removing seasonal bias from our study. Tow samples that had been stored in ethanol immediately after collecting were analysed under light microscope and picked for planktonic foraminifera, with selected specimens imaged using a scanning electron microscope (SEM) to test for any visible signs of dissolution (Fig. 2). Of these, 24 specimens (12 historical and 12 modern) were then selected for imaging using X-ray computed tomography (CT). Nano-CT scanning provides an innovative technique for assessing shell properties, and offers the advantage of yielding many different parameters (wall www.nature.com/scientificreports www.nature.com/scientificreports/ thickness, chamber volumes etc.), per single scan. It also allows a comprehensive investigation into the entire shell, not just small sections, to display changes in shell thickness during the life cycle of the protozoan, whilst also taking into account areas of differing thickness due to shell structure morphology 22 (Fig. 3).
Inspection of the outer surface of the shells using SEM revealed all selected specimens to be in good condition with no evidence of dissolution or recrystallisation 23,24 (Fig. 2). A subsample of test calcite from Globigerina bulloides was further analysed using X-ray diffraction, which revealed no differences in CaCO 3 peaks between the HMS Challenger and Tara specimens (Fig. 4).
The results of the Nano CT-scanning revealed that, without exception, all modern foraminifera specimens had measurably thinner shells than their historical counterparts (Figs. 3 and 4A-D). In addition, due to the extremely thin walls of several specimens from Tara data set, the initial voxel size of 2.4 μm was insufficient to properly render the data. A total of 4 specimens were scanned at sub-micron resolution (~0.7 um); 2 from HMS Challenger (1772-1876) samples and 2 from Tara Oceans (2009-2013) plankton tows. We analysed two species: Globigerinoides ruber and Neogloboquadrina dutertrei, and collected between 8 and 27 replicate measurements of test thickness per chamber, from the final 3 chambers of the test (Supplementary Table 1). A Shapiro Wilks normality test was performed on the chamber thickness data and an independent t-test assuming equal variance was subsequently selected to test whether the observed differences in shell thickness between preindustrial and modern specimens are significant.
Analysis of the higher resolution scans (~0.7 μm) revealed that specimens of N. dutertrei from both the historic and modern assemblages show the similar patterns of shell thickness, in that the walls of test chambers become progressively thinner towards the final chamber (Fig. 4A,B). The specimens analysed in the first scan (Fig. 4A) revealed a mean shell thickness of 28.60 ± 1.43 μm in the HMS Challenger specimen compared to 14.94 ± 0.73 μm recorded in the Tara specimen. The specimens of N. dutertrei analysed in the second scan ( Fig. 4B) record similar results with a mean shell thickness of 24.99 ± 0.85 μm in the HMS Challenger specimen, and 5.96 ± 0.23 μm in the Tara, equating to a reduction in shell thickness of up to 76%.
Quantitative measurements of the test wall in specimens of G. ruber revealed mean test thickness of 16.34 ± 0.39 μm (Fig. 4C) and 15.5 5 ± 0.51 μm (Fig. 4D) for the HMS Challenger specimens and 11.74 ± 0.53 μm and 13.96 ± 0.45 μm in the Tara specimens. This equates to an overall reduction in shell thickness of up to 30%, however statistical analysis of the individual chambers in G. ruber revealed that the observed differences in test thickness of chamber 3, between the preindustrial and the modern were not significant. Over all this species records similar but less marked differences in the mean shell thickness compared to N. dutertrei. This may be due to differences in calcification between species.

Discussion
Initial findings of the Challenger Revisited project have revealed a reduction in shell thickness of up to 76% in selected species of planktonic foraminifera over the last c. 140 years, which corresponds to a period of profound change in our oceans. Ocean acidification is not the only stressor faced by the world's oceans in the coming decades and over the time period studied here. Rising temperatures and deoxygenation are also likely to have a substantial impact on marine ecosystems, and eastern boundary upwelling systems are likely to be strongly affected by all three stressors 20   www.nature.com/scientificreports www.nature.com/scientificreports/ measurements (1872-1876) to a more recent data set from the Argo Programme (2004-2010) revealed substantial warming of the modern upper ocean, with a warming signal that is global in its extent. We propose OA as the lead cause for the observed reduction in calcification of planktonic foraminifera as our results closely mirror results from laboratory studies and field observations from the Arabian sea 17 . However, It could be argued that rising temperatures, deoxygenation and OA are intrinsically linked and therefore these stressors should not be separated. More work is required to further explore the application of this method by increasing sample sizes Pink circles represent HMS Challenger (1875) specimens, blue circles represent Tara (2011) specimens, squares represent the mean thickness for each chamber; (E) X-ray diffraction data for specimens of G. bulloides. Black line represents HMS Challenger data, red line represents Tara data. (2020) 10:1620 | https://doi.org/10.1038/s41598-020-58501-w www.nature.com/scientificreports www.nature.com/scientificreports/ and undertaking comparative studies in other ocean basins. Though the HMS Challenger samples only provide a snapshot of early industrial ocean conditions, the method has potential for global reconstructions. The "Challenger Revisited project" will be expanded beyond the Pacific Ocean, and the methodology can be applied to other historical expedition collections such as the Terra Nova [1911][1912][1913][1914][1915] and Discovery [1920-1930s] when comparatory contemporary expedition materials are available.
Whilst all specimens analysed showed some reduction in shell thickness, the degree to which different species responded varied greatly. Specimens of N. dutertrei, a non-spinose, thermocline-dwelling planktic foraminifera that possesses intracellular chrysophyte alga 25,26 revealed up to 76% reduction in shell thickness between the preindustrial and the modern, whereas G. ruber specimens display a far smaller decrease in shell thickness (~20%). G. ruber is a spinose multichambered species known to occupy the mixed layer of the ocean and hosts photosynthesising algal symbionts (dinoflagellates) which can alter the chemistry of the sea water immediately surrounding the shell and therefore enhance calcification 27,28 . Numerous studies have demonstrated that a variety of calcifying organisms respond negatively to decreasing ocean pH, such as coccolithophores, pteropods and corals 22,29,30 . However, certain photosynthesising organisms have been shown to benefit from higher availability of dissolved CO 2 31 .
To further investigate all drivers for these differences in shell thickness, the effect of depth habitat on calcification ability is worth further study, as is how photosymbiont hosting foraminifera and also the types of symbionts may be more resilient to decreasing pH. Studies of the fossil record have shown that foraminifera, when under extreme stress, will shed their symbionts in the process of bleaching 32 .
In addition, the application of directed geochemical analyses have the potential to disentangle the multiple factors that could be driving the reductions in shell thicknesses shown here. Boron isotope analysis on comparable historic and modern planktonic foraminifera shells can provide accurate reconstructions of pH conditions at the time of shell development [33][34][35] , which in turn links the reduction of ocean carbonate ions as the driver for reduced shell calcification. There is therefore the potential to gain new insights into the rate of change taking place in the oceans, contributing greatly to our understanding of the sensitivity of global climate systems to CO 2 forcing. Such insights are crucial for predicting the future climate and health of the oceans.

Sample Information
Our preindustrial specimens were collected at HMS Challenger station 272 in the Southwest Pacific Ocean on the 8 th of September 1875 (3°48′0″S, 152°56′0″W), using a tow net "resembling long night-caps, of fine muslin or calico, and 10 to 16 inches in diameter at the mouth" 1 . In the HMS Challenger volumes the nets are reported as being towed at various depths, "even as far as 800 fathoms" 36 , however the ship's station book does not record the depth at which the nets were towed at individual stations. Sea surface and bottom temperatures, the sounding, "nature of the sea floor" and species recovered are recorded. The expedition logs are publicly available at the library of the Natural History Museum, London.
A small sub-sample of the HMS Challenger material (50 g) was taken from the stores and gently washed using deionised water, the residues were dried in an oven at 40 o C. Specimens were selected from the 250-500 um size fraction for further analysis.
The Tara Oceans specimens came from station 127 (6°37′12.00″S, 152°34′4.80″W) and station 128 (0°28′8.04″S, 153°18′19.44″W) in the Pacific ocean and were collected between the 31 st of August and 5 th of September 2011 from the Deep Chlorophyll Maximum (DCM) layer (100 m) 21 . We analysed 10 ml aliquots from the 180-2000 um size fraction, samples were stored in a freezer in ethanol. Detailed accounts of the collection methodology can be found in Pesant et al. 2015.
Species identifications follow the existing understanding of modern foraminiferal taxonomy [37][38][39] . The primary classification is based on the chamber arrangement, wall structure and principally spinose or non-spinose ornamentation 40,41 .

Methods
Nano-CT scanning. Nano-CT scanning was carried out with a Zeiss Versa 520. Samples were analysed in pairs using a customised pine wood mount. Sample mount, X-ray source and detector geometry were kept constant throughout the first set of scans (Fig. 2). A scan resolution voxel (a 3D unit of space which varies in dimensions between CT reconstructions depending on scanning parameters) size of 2.4 µm 3 was typically achieved using this set up in order to maximize the number of specimens that could be analysed in a single scan. During the second run of scans (Fig. 3) focusing on just 4 specimens (2 historic and 2 modern), X-ray source and detector were positioned to maximize resolution and a voxel size of <0.8 µm 3 was achieved. Between 700-1000 individual X-ray absorption profiles of each mount were taken and combined to build a 3D rendering of the image using Avizo 9.2 software which was also used for image analysis.
The differential X-ray absorption of organic matter, pine wood, and calcium carbonate translates into different greyscale intensities in the reconstructed 3D image; therefore, it was possible to filter out pine wood and organic material from the analysed data for volume, wall thickness etc. Object volume, average thickness and surface area were determined using standard Volume Graphics analysis. In addition, dimension measurements were taken on 3D renderings of shells. For N. dutertrei and G. ruber, three dimensional measurements were taken, the width (at the widest point), the width at half shell length and length (Fig. 2a). Object volume, average thickness and surface area were determined using standard Volume Graphics analysis. For direct comparisons between shell thickness, weight and density of the old and new samples, a size filter was applied to the dataset and only individuals of the same length were used. X-ray diffraction. Specimens of Globigerina bulloides were picked for X-ray diffraction analysis as they were relatively abundant in both Tara and Challenger samples, between 5 to 10 shells were soaked in ethanol and Scientific RepoRtS | (2020) 10:1620 | https://doi.org/10.1038/s41598-020-58501-w www.nature.com/scientificreports www.nature.com/scientificreports/ gently ground in an agate mortar and deposited on a circular sapphire substrate. The XRD data were collected using an ENRAF-Nonius 590 diffractometer with a Tintel curved position sensitive detector (PSD). The angular linearity of the PSD was calibrated with silicone powder and silver behenate. This apparatus collects data from 2-120 °2θ continuously throughout the experiment. Cobalt Ka radiation was selected from the primary beam by a Germanium 111 crystal monochromator with the x-ray tube operating at 40 kV and 30 mA. Horizontal and vertical slits restricted the beam to a height of 0.14 mm and width of 5 mm. Data were collected with samples spinning continuously in the plane of the sample surface and with the sample surface at an angle of 4° to the incident beam.
Detailed statistics. Neogloboquadrina dutertrei. Scan 1. In chamber 1 (the final chamber in the whorl), the preindustrial (Challenger) specimens recorded a mean shell thickness of 20.60 ± 0.56 μm, and a mean thickness of 8.93 ± 0.44 μm was recorded in the modern (Tara) specimen. A two sample t-test assuming equal variance was carried out. The t-statistic calculated was 16.593 and the two-tailed probability was 0.000. Therefore we can conclude there is a significant difference in shell thickness between the Challenger and Tara specimens.
In chamber 2, the preindustrial (Challenger) specimens recorded a mean shell thickness of 28.02 ± 1.2 μm, and a mean thickness of 18.12 ± 0.89 μm was recorded in the modern (Tara) specimen. A two sample t-test assuming equal variance was carried out. The t-statistic calculated was 6.71 and the two-tailed probability was 0.000. Therefore we can conclude there is a significant difference in shell thickness between the Challenger and Tara specimens.
In chamber 3, the preindustrial (Challenger) specimens recorded a mean shell thickness of 37.98 ± 1.30 μm, and a mean thickness of 16.19 ± 0.90 μm was recorded in the modern (Tara) specimen. A Mann Whitney U test showed that there was a significant difference (U = 170, p = 0.000) between the preindustrial and modern specimens. Scan 2. In chamber 1 (the final chamber in the whorl), the preindustrial (Challenger) specimens recorded a mean shell thickness of 20.16 ± 0.65 μm, and a mean thickness of 4.33 ± 0.17 μm was recorded in the modern (Tara) specimen. A two sample t-test assuming unequal variance was carried out. The t-statistic calculated was 24.142 and the two-tailed probability was 0.000. Therefore we can conclude there is a significant difference in shell thickness between the Challenger and Tara specimens.
In chamber 2, the preindustrial (Challenger) specimens recorded a mean shell thickness of 22.55 ± 0.59 μm, and a mean thickness of 5.44 ± 0.27 μm was recorded in the modern (Tara) specimen. A two sample t-test assuming unequal variance was carried out. The t-statistic calculated was 26.521 and the two-tailed probability was 0.000. Therefore we can conclude there is a significant difference in shell thickness between the Challenger and Tara specimens.
In chamber 3, the preindustrial (Challenger) specimens recorded a mean shell thickness of 32.87 ± 0.90 μm, and a mean thickness of 7.89 ± 0.33 μm was recorded in the modern (Tara) specimen. A Mann Whitney U test showed that there was a significant difference (U = 432, p = 0.000) between the preindustrial and modern specimens.
Globigerinoides ruber. Scan 1 In chamber 1 (the final chamber in the whorl), the preindustrial (Challenger) specimens recorded a mean shell thickness of 16.60 ± 0.59 μm, and a mean thickness of 10.38 ± 0.59 μm was recorded in the modern (Tara) specimen. A two sample t-test assuming equal variance was carried out. The t-statistic calculated was -7.459 and the two-tailed probability was 0.000. Therefore we can conclude there is a significant difference in shell thickness between the Challenger and Tara specimens.
In chamber 2, the preindustrial (Challenger) specimens recorded a mean shell thickness of 15.74 ± 0.74 μm, and a mean thickness of 12.74 ± 1.12 μm was recorded in the modern (Tara) specimen A Mann Whitney U test showed that there was a not a significant difference (U = 95.5, p = 0.000) between the preindustrial and modern specimens.
In chamber 3, the preindustrial (Challenger) specimens recorded a mean shell thickness of 17.40 ± 0.78 μm, and a mean thickness of 14.52 ± 0.45 μm was recorded in the modern (Tara) specimen. However insufficient replicate measurements were collected (<10) to perform statistical analysis of the data. Scan 2. In chamber 1, the preindustrial (Challenger) specimens recorded a mean shell thickness of 12.71 ± 0.25 μm, and a mean thickness of 10.01 ± 0.47 μm was recorded in the modern (Tara) specimen. A Mann Whitney U test showed that there was a significant difference (U = 58.5, p = 0.000) between the preindustrial and modern specimens.
In chamber 2, the preindustrial (Challenger) specimens recorded a mean shell thickness of 18.77 ± 0.64 μm, and a mean thickness of 15.06 ± 0.48 μm was recorded in the modern (Tara) specimen. A two sample t-test assuming equal variance was carried out. The t-statistic calculated was 4.749 and the two-tailed probability was 0.000. Therefore we can conclude there is a significant difference in shell thickness between the Challenger and Tara specimens.
Statistical analysis of the measurements taken on chamber 3 suggest that any observed difference in shell thickness is not significant.