× Dismiss

Never Miss an Update.

News Deeply will use the information you provide to send you newsletter updates and other announcements. See our privacy policy for more.

Oceans Deeply is designed to help you understand the complex web of environmental, social and economic issues facing the world’s oceans. Our editors and expert contributors work to bring you greater clarity and comprehensive coverage of ocean health.

Sign up for our newsletter to receive weekly updates, special reports and featured insights as we cover some of the most critical issues of our time.

To Save the Ocean Now, Museum Aims to Map Species of the Past

As New York’s American Museum of Natural History digitizes its vast collections, it wants to use that historic data to help scientists understand how imperiled ecosystems are changing today. But coders at a recent hackathon discovered just how big a challenge that is.

Written by Jessica Leber Published on Read time Approx. 4 minutes
Coders work together at the Hack the Deep event.©AMNH/C. Chesek

Today, visitors to the Bahamas’ Andros coral reef, the third-largest barrier reef system in the world, will see it as a shell of what it once was. Scientists have categorized more than half of the reef as in “poor” or “impaired” health, according to a 2016 report, as it faces threats such as climate change and illegal fishing. The International Union for Conservation of Nature estimated in 2012 that all but 10 percent of reefs throughout the Caribbean are dead or dying, even before the major bleaching events of recent years.

None of this is what marine biologist Roy W. Miner must have experienced when he conducted five expeditions to the Andros Reef and the Bahamas in the 1920s and early 1930s. His mission was to collect coral specimens as a curator for the American Museum of Natural History (AMNH) in New York City, one of the largest natural history museums in the world. (Watch a fascinating video of Miner’s activities here.)

His work provides part of the museum’s 33 million-plus specimens and artifacts, including around 8 million marine invertebrates, with some corals dating back to 1864 and a rare giant squid. These are the kinds of historic collections that are valuable to scientists seeking to understand the unprecedented and rapid destruction of marine ecosystems around the world today.

“Our coral collection isn’t particularly large … but it’s pretty diverse and old. That’s the value,” said Christine Johnson, a curatorial associate at the American Museum of Natural History. If you went back to these same areas today, she said, it would be useful to “see if the species distribution was the same 100 years ago.”

Around the world, over the past decade or so, many natural history museums and research institutions have been embarking on massive efforts to digitize older collections that have long sat tagged in drawers, in part to make such information more readily available at a time when historic context about the world’s biodiversity is sorely needed.

image11.jpg

Example of a map and search tool created by the team GS@HacktheDeep during the event.

From the National Museum of Natural History in Washington, D.C., working to digitize and image its plant collections as rainforests are destroyed, to the grant that the AMNH received in 2016 to preserve, organize and image its 4,000-strong dry coral collection, this work is the most urgent for vulnerable ecosystems.

As more marine data becomes digitized, the AMNH explored what it could do with it at its fourth annual hackathon, called Hack the Deep, held at the museum last Sunday, which invited coders to help visualize the museum’s oceanic data.

One of several goals at the event was to visualize specimen data on maps to assist scientists who might want to understand what marine communities or ecosystems might have looked like in the past or how they’ve changed over time. For Johnson, this kind of thinking takes inspiration from related local research endeavors, such as the Wildlife Conservation Society’s Welikia Project, which is working to visualize New York City’s native ecology before European settlers arrived.

The value of doing this kind of work is illustrated by studies such as one conducted on the tidal pools of California’s Monterey Bay, according to Melanie Hopkins, an assistant curator in the AMNH’s invertebrate paleontology division. In the 1990s, scientists did population and biodiversity surveys in the same areas as were rigorously documented in the 1930s and were able to gauge the decline in many species there as waters in the bay had warmed.

“If we can do this on a global scale and on a collections scale, we can see a much broader perspective of those habitat shifts – not just one little place,” said Hopkins, who was among the hackathon organizers. (In the case of Monterey Bay, it helped that the brass bolts used to mark survey areas in the 1930s were still in place.)

One group, Team L, tried working to group and map data by eco-region.

However, the difficulty in even getting the needed information in place was apparent at the weekend’s hackathon, where more than 150 coders gathered for an intense overnight visit under the museum’s famous 94ft (29m) “blue whale” ceiling display. They worked on nine projects related to ocean science and conservation, such as using computer vision and machine learning algorithms to identify and count organisms like sea anemones or marine fossil trilobites and creating visual 3D models of microscopic marine organisms.

However, by far the biggest challenge, both in terms of the number of coders needed and its scope, was the one set up by Johnson to create maps of the museum’s digitized collections. Coders were tasked with taking a sample data set of 75,000 marine specimen records, setting up scripts to “clean” the data so it would be in a usable database format, attaching geographic coordinates and finally putting the results on a map that could show snapshots over time.

“In museums, we tend to separate our specimens – corals go over there, sponges go over there – the information is dissociated and modular,” said Johnson. “It’s very hard to put the community back together if it’s not in any digitized form.”

In addition to being limited to a single day, the five teams working on the challenge encountered plenty of stumbling blocks, including incomplete records, and misspellings of names and locations in the data. In their final presentations, one team explained the hard decisions that had to be made in even visualizing the data on a map when some sample labels were very specific about exactly where they were collected and others might just have said something like “India.”

As in almost every hackathon, it was clear that a lot more work was needed than could be accomplished in just one day. The museum hopes that this hackathon will spawn further collaborations, Hopkins said, as they have in past events the museum has held. As the museum’s digital images and records of collections increase in number, there will be more research opportunities to mine the data.

“We’re starting to get to a place where we actually have large enough collections of images that we can take advantage of some of these semi-automated tools to collect a lot more data, a lot faster,” Hopkins said.

Some content here

× Dismiss
We have updated our Privacy Policy with a few important changes specific to General Data Protection Regulations (GDPR) and our use of cookies. If you continue to use this site, you consent to our use of cookies. Read our full Privacy Policy here.

Become a Contributor.

Have a story idea? Interested in adding your voice to our growing community?

Learn more