In my last post, I discussed some of the sessions I attended at the Digital Data in Biodiversity Research Conference held at the University of Michigan and sponsored by the University with the iDigBio project. More than half the records in the iDigBio portal are for plant specimens, and as several speakers noted, plants are easy: most specimens are relatively flat and can be scanned or photographed much as the pages of a book are. Bones are different. A 2D image leaves a lot to be desired, but 3D images are expensive to produce and also to store since the file size is so large. Daniel Fisher of the Museum of Paleontology at the University of Michigan discussed 3D surface models of mastodon bones where images have been converted into replicas. The mastodon in question was unearthed at an Indiana site and the replica is now on display at the Museum of Paleontology, so it’s a nice example of home-grown expertise. Fisher also referred to the University of Michigan Online Repository of Fossils, where even a plant-lover can become mesmerized by bones, some in 3D—and there are botanical fossils as well.
Mark Uhen continued the emphasis on large mammals with a presentation on fossil whales. However, they only take up a small part of the data portal he was instrumental in creating the Paleobiology Database (PBDB) and its well-developed user platform PBDB Navigator. His goal for the site is to document every fossil occurrence on the planet; this is the idea of a portal writ large. It’s an attractive though farsighted goal, since only a minority of fossils are databased at the moment, and the issue of data heterogeneity means that even this portion couldn’t just be loaded into a single portal. Several speakers spoke of such ambitious goals, ones that might seem like pipedreams, but with the rate at which the digitization field is progressing thanks to advances in data storage, standardizing metadata, and relying evermore on automation, looking forward is crucial to moving forward.
Still another portal, MorphoSource, is a virtual museum and digital repository for 3D specimen data. As the presenter Doug Boyer of Duke University noted: “I think there is even some plant data” in it, suggesting how extensive and democratic the portal is. Then Susan Butts of Yale University discussed a project called ePANDDA (Enhancing Paleontological and Neontological Data Discovery API). Since it deals with making data accessible to several audiences, its developers wanted an appealing acronym (yes, the logo includes a panda’s face). The project aims at linking information in iDigBio with that in PDBD and iDigPaleo. As Butts argues, there isn’t a need for a new database but for integration of databases. That was also the theme of Duke University’s Julie Winchester who discussed the challenges of integrating MorphoSource and iDigBio. I have been to a couple of iDigBio meetings in the past, and in comparison there was more emphasis here on integration, and also on larger datasets, such as those involved with 3D imaging.
This trend was also apparent in other presentations. Adam Summers of the University of Washington discussed his project on 3D scanning fish skeletons in which he scans a pile of fish—ten or more at a time—to make the process more efficient and economical. This approach obviously involves some post-scan manipulation of datasets, but the unambiguous images that result speak for themselves. Beth Brainerd of Brown University deals with even larger datasets using XROMM, X-ray Reconstruction of Moving Morphology. XROMM combines 3D models of bone morphology with movement data from biplanar x-ray videos to create highly accurate animations of bones moving in 3D space. This enables an analysis of complex processes like a bird flying and a frog jumping. Surangi Punyasena of the University of Illinois discussed 3D imaging of pollen grains to provide information on the extinction of the spruce Picea crutchfieldii during the last ice age. Detailed scans were needed to differentiate this species’ pollen from that of two other spruce species that were also present at the time. The project was interesting to me because it involved plants (Alleluia!) and very sophisticated analysis not only to detect small differences but also to correct for distortion in fossils.
After all these presentations, it was obvious why this conference was so necessary as a way to alert the natural history collection community to the direction digitization will be taking in the future. The emphasis was definitely on imaging as much as on other forms of data, and thus on large files that need to be stored, integrated with other data, and presented on portals that are both accessible and easy to use. Some solutions were suggested, particularly for paleontological databases and efforts to integrate them. Also, the pollen and biomechanics projects presented ways that databases could also be employed to manage research projects and further communication among researchers. It will be very interesting to see what’s reported at future iDigBio events. With all this data, it’s important to remember what’s at the root of iDigBio: natural history collections all over the United States. During my road trip, I got to visit two plant collections that are indicative of the riches available there. These will be the subject of my next post.