Kelp Wanted: Segmenting Kelp Forests

Help researchers estimate the extent of Giant Kelp Forests by segmenting Landsat imagery. #climate

$15,000 in prizes
feb 2024
671 joined

About the data

Giant Kelp

Most life on the seafloor can only be sampled by SCUBA divers or dredging up samples from the deep. This kind of data requires a ton of (really fun) effort to collect, but it means scientists are limited to places they can get to! But there are some vitally important species in the sea that be seen from as far away as space, alleviating the need to slip on some fins to collect data. The data presented in this challenge represents satellite observations of the vitally important giant kelp, Macrocystis pyrifera.

Overhead drone footage of giant kelp canopy. Image courtesy of Tom Bell, All Rights Reserved.

Overhead drone footage of giant kelp canopy. Image Credit: Tom Bell, All Rights Reserved.

Giant kelp is amazing - it can grow up to a foot a day and forms lush canopies that can be seen by some of the earliest satellites humankind put into space! Kelp has captured people's attention for centuries. Its high iodine content saw it used as a medieval goiter treatment. Darwin wrote about it in The Voyage of the Beagle. It was burned to ash and used to make glass in the 1700s and 1800s. These days, we recognize that kelp is what we call a ‘foundation species’. In ecosystems where kelps are present (roughly 25% of the world’s coastlines!), it forms the foundation of the entire ecosystem. It provides food for all manner of herbivores from tiny shrimp to ravenous sea urchins to grazing fish. Giant kelp is extraordinary in that, of all of the kelps, it has one of the widest global distributions and is one of the easiest to see from satellites.

Labeling Landsat with Citizen Scientists

The imagery in this dataset comes from the Landsat series of satellites. These satellites photograph the entire surface of the earth every 16 days, and the earliest Landsat satellites used in this competition have been doing so since 1984. When scientists first began working with these images, they realized swiftly that classification was going to be a difficult task. Landsat was not designed to be able to see kelp, but kelp’s reflectance signature (the color of light that it reflects) is just within the Landsat camera's detection abilities. Because of this, kelp and something as simple as the glint of sun off of a wave may look the same to a computer.

To solve this problem, kelp researchers have joined forces with citizen scientists via the website Floating Forests, built in collaboration with Zooniverse. Citizen scientists use the site to classify kelp from Landsat imagery that were reprocessed to be visually easy to use. They’ve generated tens of thousands of classifications over multiple project rounds. This has been great for specific projects, but is still a painstaking process. It also does not provide data quickly enough for scientists to take a year of Landsat imagery and easily whip up an annual global map of giant kelp to determine if the environment has caused widespread change in the species, which is important information for stakeholders from environmental policymakers to kelp stewards.

A Global Kelp Model

Through this competition, scientists from UMass Boston and the Woods Hole Oceanographic Institution are hoping to find a model that can be used around the planet to swiftly and efficiently classify giant kelp with at least at the same accuracy as the human eye. The provided Landsat imagery is split up into tiles that match those shown on Floating Forests in hopes that competition participants will be able to create such a model. The uses and needs for such a model are great, and it will doubtlessly allow for some incredible leaps in the science of global change for our oceans. We hope you enjoy the challenge!

About the partners

The Zooniverse is the world’s largest and most popular platform for people-powered research. This research is made possible by volunteers — more than a million people around the world who come together to assist professional researchers. Zooniverse's goal is to enable research that would not be possible, or practical, otherwise. Zooniverse research results in new discoveries, datasets useful to the wider research community, and many publications.

Woods Hole Oceanographic Institution (WHOI) is the world’s leading, independent non-profit organization dedicated to advancing ocean research, exploration, and education.

Tom Bell is an Assistant Scientist in the Department of Applied Ocean Physics and Engineering at WHOI. His research uses a variety of remote sensing platforms including satellites, drones, and occupied aerial imagery to investigate the abundance and physiological condition dynamics of kelp forests, and rocky intertidal and salt marsh ecosystems across local and global scales.

Henry Houskeeper is a postdoctoral scholar at WHOI interested in better understanding how ocean environments, such as kelp forests, respond to changes including rising temperatures and shifting currents. Henry conducts research studying the light illumination of underwater environments and exploring how best to use images from satellites to understand and monitor marine environments. When not conducting marine research, Henry can be found swimming or enjoying the beaches of Cape Cod.

Jarrett Byrnes is an associate professor of Biology at the University of Massachusetts Boston broadly interested in the causes and consequences of changes in kelp forests around the globe. He has been the PI on two NASA grants to fund the development of Floating Forests and connect community scientists with kelp forests through the medium of online projects. When not coding away with remote sensing kelp forest data, he can be found underwater studying the vast meadows of kelp found in the Gulf of Maine.

You can learn more here

An image of the Woods Hole Oceanographic Institution logo          An image of the UMass Boston logo

An image of Zooniverse