To verify that all microfossils were identified and selected from each slide, we visually checked the boxed slide output. Image selection parameters in segment were adjusted as needed to optimize object selection. For a given set of image segmentation parameters, object selection is deterministic i.
The deterministic nature of the object selection software was verified by re-segmenting three slides twice and one slide three times and confirming the number and identity of objects. The number of objects outputted by segment were then cross checked with the number of objects outputted by all following modules focus , run2dmorph , and run3dmorph for each slide. Both run2dmorph and run3dmorph output lists of objects with failed outline or mesh extractions. These lists were examined for each slide to ensure that complete foraminifera were included and that specific species were not being disproportionately missed.
When problematic e. The same set of image extraction parameters yielded satisfactory results for 3D shape extractions of complete planktonic foraminifera from all samples.
The accuracy and reproducibility of 2D and 3D size extraction was confirmed with direct measurements. For run2dmorph , a calibration slide IP. In total, ten complete planktonic foraminifera from four species were measured along their minor and major axes using a stage micrometer on a Leica S8APO microscope. The calibration slide was also segmented with each of the three code versions of the segment module of AutoMorph , and then processed through run2dmorph to obtain automated measurements of the major and minor axis for each individual foraminifer.
To do this, the ImageJ scale was set using the automatic scale bar added to the image label by segment , and the major and minor axes were drawn by hand. The three measurement types run2dmorph , ImageJ and stage micrometer were then compared Table 4 and Fig. In both panels, object measurements are normalized to the mean measurement to highlight the variation between repeated measurements and the relative reproducibility of both approaches.
AutoMorph Fig. Repeated hand measurements in ImageJ had as much as a 16 micron difference between measurements. Importantly, all three approaches AutoMorph , ImageJ , and measurement with a stage micrometer provide the same average 2D measurements for foraminifera Fig. Together, these tests indicate that AutoMorph provides accurate and reproducible 2D measurements of foraminifera. The accuracy and precision of 3D size extraction was previously assessed 7 by comparing the height extraction with the length and width of spherical objects and by examining the effect of object orientation and imaging conditions on 3D mesh extraction and volume estimation see ref.
These tests indicated height extraction within 7. Repeated measurements of ten specimens listed in Table 4 using three versions of AutoMorph a or three repeated hand measurements in ImageJ b. Individual-specific data in a and b plotted as residuals of the mean individual-specific size measurement. Each individual labelled in a is represented by two columns of data major axis and minor axis.
Unter den Schollen tobt das Leben
Regression of micrometer measurements black squares and average ImageJ measurements blue crosses as a function of average AutoMorph size c did not significantly differ from the line light grey line. Extensive spot checks of final EDF image classifications found object classification by human observers to be The errors are described briefly here category listed in quotes followed by a list of object-types included in error , with each classification category described in more detail in Usage Notes. Chunks of consolidated sediment were generally poorly classified.
These combinations were assigned the classification of the larger object in cases where the small foraminifera were completely nested within the outline of the larger object. The splits of core top samples used in this study were, to our knowledge, unbiased by previous research efforts undertaken on the material, with exception to the benthic foraminifera.
Many of the samples were picked for specific species of benthic foraminifera in the past, so all benthic foraminifera results should be considered as illustrative of some of the species present but not necessarily quantitative representations of their original abundance or full diversity in the samples. More generally, it is worth noting that most of the core top samples used here have a long collection history in other laboratories, so it is possible that selective sampling of some planktonic foraminifera or other species occurred in the past without our knowledge.
Besides this effect, it is worth reiterating that the assemblage data provided here comes from death assemblages. In spite of visual evidence for good preservation in most of the core top samples included, selective dissolution of small-bodied and delicate species is known to begin even in the water column 20 , 21 , and the assemblages imaged are certainly time-averaged on the scale of hundreds to many thousands of years. Although we include all images extracted by segment in this dataset, do note that our initial sieve size was microns.
Foraminifera - Diatom Shop
Although there are a number of objects smaller than microns in this dataset, they are not representative of the abundance of this size category in the original sample. Rather, they are the rare objects that slipped through our size filter, and should be excluded for most applications. At least one ancient fossil appears in the core top data set. We have left this ancient fossil in as an indication of the level of cross-contamination in the lab very low but potentially present. It is also possible that this stratigraphically out of place foraminifera was reworked in the sediments or introduced during handling in other labs.
Regardless, users should remove this such outliers in species-specific applications. Here we described the post-processing that we carried out to insure that images from these samples accurately reflect species and size distributions at those sites. This size-fractionated handling of these sites Sites IPE. To correct for this bias, it was necessary to subsample the object output from these slides to properly represent the relative distributions of objects in the original sample.
More specifically, for IPE. This data report includes objects after down-sampling and should be corrected for the bias introduced during slide preparation. In a few classification categories, the confidence categories were used to indicate other attributes; these exceptions are explained below. Classification categories, listed in Fig. Low confidence in this category i.
Lower to low confidence i. Damaged tests of planktonic foraminifera were classified as 'damaged' for all breaks, drill-holes, and damage assessed to affect less than around a third of the test. In each of these diatom, echinoid spine, mollusk, ostracod, and radiolarian , complete or large fragments of organisms were typically identified with greater confidence than small or out-of-focus pieces.
Echinoid spines were confirmed as echinoid in nature by the match of the distinctive lattice structure in spine images with those of an immature echinoid in the YPM Invertebrate Zoology collection YPM IZ. The 'unknown' category contains non-target items, such as bits of background from the slide, fibres, and other unknown objects. Small pebbles, minerals, and other rock-like objects were categorized as 'rock'.
Objects in direct or very near contact cannot be accurately extracted for 2D and 3D morphometrics. How to cite this article : Elder, L. Sixty-one thousand recent planktonic foraminifera from the Atlantic Ocean. Data doi: Sutton, M. Tomographic techniques for the study of exceptionally preserved fossils.
B , — Cunningham, J. A virtual world of paleontology. Trend Ecol. Friis, E. Phase-contrast x-ray microtomography links cretaceous seeds with Gnetales and Bennettitales. Nature , —U Schmidt, D. Linking evolution and development: Synchrotron Radiation X-ray tomographic microscopy of planktic foraminifers.
Palaeontology 56 , — Siveter, D. Current Biology 24 , — Hsiang, A.
Recent Benthonic Foraminifera: Environmental Factors controlling their Distribution
AutoMorph: accelerating morphometrics with automated 2D and 3D image processing and shape extraction. Methods Ecol. Towards a morphological metric of assemblage dynamics in the fossil record: a test case using planktonic foraminifera. B , Hemleben, C. Modern Planktonic Foraminifera Springer-Verlag , Darling, K. The genetic diversity of planktic foraminifera and the global distribution of ribosomal RNA genotypes. Beaufort, L. Automatic recognition of coccoliths by dynamical neural networks.
Optical measurements to determine the thickness of calcite crystals and the mass of thin carbonate particles such as coccoliths. Nature Protocols 9 , — Bollmann, J. Francus Ch. Abiotic forcing of plankton evolution in the Cenozoic. Science , — Rowe, T. The Disappearing Third Dimension. Schiebel, R. Modern planktic foraminifera. Molecular evidence of cryptic speciation in planktonic foraminifers and their relation to oceanic provinces. USA 96 , —, Archer, D. An atlas of the distribution of calcium carbonate in sediments of the deep sea.
- Sex Differences: Developmental and Evolutionary Strategies.
- The Complete Tolkien Companion.
- Alien Dawn: A Classic Investigation into the Contact Experience.
- Sixty-one thousand recent planktonic foraminifera from the Atlantic Ocean.
Global Biogeochem. Ridgwell, A. Regulation of atmospheric CO2 by deep-sea sediments in an Earth system model. Berger, W. Foraminiferal ooze - solution at depths. Deep-Sea Res. Polly, P. History matters: ecometrics and integrative climate change biology.
- Swipe to navigate through the chapters of this book!
- What are Foraminifera?!
- Unwept (The Nightbirds, Book 1).
- Recent foraminifera in the Arctic Ocean and Kara Sea continental margin.
- Beginning Database Design Solutions.
- ArcGIS Web Development?
Sun, X. The effect of primary productivity and seasonality on the distribution of deep-sea benthic foraminifera in the North Atlantic. Download references.
Elder, L. E et al. We would like to thank Joanne Zhenheng Li, Casey Culligan and Emma Tipton for assistance in arranging slides, Liana Epstein for assistance in testing the image classification app, and Gregory Meyer for contributing code on the classification app.
LEE performed all imaging and technical validations. All objects were classified by P. Classification app was developed by A. Correspondence to Pincelli M. Paleoceanography and Paleoclimatology Article metrics. Advanced search. Skip to main content. Muthuswami Memorial Lecture C. Vaidyanadhan Award Endowment Lecture C. Mahadevan Endowment Lecture R. Prahalad Rao Memorial Lecture Dr. Wadia Endowment Lecture G. Sawkar endowment for Teachers training V. Future Challenges in Earth User Username. Remember me. Article Tools Print this article. Indexing metadata.
How to cite item. Finding References.