Breakthroughs in Science: The Global Fishing Watch Research Program

Breakthroughs in Science: The Global Fishing Watch Research Program

By | 2019-08-23T11:05:21+00:00 August 21st, 2019|Data and technology, Fisheries, Research and analysis|

Why is Global Fishing Watch (GFW), a non-profit organization, investing so much effort into collaborating with scientists to publish research papers? And why has the program been so successful in doing so? In this blog, GFW Research and Innovation Director, David Kroodsma outlines why we are pursuing this work, why he believes it has been successful, and what is next for the Global Fishing Watch research program.

The key reason we work with the scientific community is that many of the biggest challenges facing the oceans are exacerbated by a complete lack of information. One third of fisheries are overfished, and the proportion has been growing over the past few decades. About 20% of fish is caught through means that are illegal, unregulated or unreported. Slave labor is commonly found on some fleets, and actual catch in the world may be more than 30% higher than the official reported catch. By better monitoring where and when fishing is happening, it will become more difficult for these activities to continue. Scientists can help us develop the needed accurate and credible datasets to bridge this information gap.

Indeed, working with scientists over the past three years, we’ve been able to facilitate discoveries that improve how we manage the oceans. Our research partners at National Geographic Pristine Seas have used our data to help create seven marine protected areas with a combined area larger than Egypt. Analyses of marine protected areas in Europe with our data has sparked a debate about ways to improve them. GFW is now being used by the World Trade Organization to help negotiate how to change more than 20 billion dollars in fisheries subsidies, many of which are encouraging overfishing. Our analysis of transshipment activity is now informing management organizations across the globe, including patrols by the U.S. Coast Guard. And our core dataset, published in Science last year, has now been downloaded over a thousand times to facilitate research and management.

Global Fishing Activity dark

“Tracking the global footprint of fisheries” was the first global assessment of fishing activity using AIS data and found that over half the ocean was fished by large industrial vessels.

This success has also resulted in a series of research papers. Over 20 have been published by our partners, including three in Science and one in Nature, and many more are in review or soon to be submitted. Last year Science Advances had a special issue on the science of the high seas, and half of the papers were by GFW research partners using our data. And in the past few months, the research program has been honored with an award for this science

We believe this success rests on a few factors: harnessing the technological revolution in satellites and big data; developing close partnerships with a network of researchers; and our commitment to open data.

The primary reason for these breakthroughs is new technology. Our key advance has been to obtain and process all of the GPS positions of fishing vessels carrying the Automatic Identification System (AIS). AIS is a device carried by large, ocean-going vessels that allows them to share their locations with nearby vessels to avoid collisions. Until recently, it was not possible to easily obtain a database of these positions. But in the past five years, more than 50 satellites have been launched that can record these messages, providing us a global database. Moreover, the same types of advances that allow Google and Facebook to process images and recognize faces allows us to process these GPS positions and identify the type of fishing vessel and when it is fishing. We have been able to harness this new machine learning technology to process this new dataset, and the result is a dramatic leap forward in our ability to monitor and understand fishing.

Another key to the research program is the nature of our collaborations. One thing we recognized early on is that the problems in the ocean are too big and complex for a team of data scientists and engineers to tackle on our own; we needed to find the experts in marine biology, economics and policy. When we started talking with these experts, we soon found that they did not have in-house resources or expertise in big data processing. In other words, the people who knew what questions to ask about ocean science and how to answer them often don’t have access to the latest machine learning technology. So we established partnerships with leading researchers (more than 10 institutions), and started working together. GFW provided the big data processing and machine learning, and the researchers provided the deep knowledge of marine science. Such partnerships are critical, as no one group has all the expertise to apply these new big data technologies.

GFW research partners workshop, hosted by Google in San Francisco, 2018.

It is also important that our results and science are open and shared widely. We strive to make as much of our data and code freely available. We also allow many researchers, including ones not officially part of our program, to access some our datasets-in-development before they are published. Surprisingly, open data isn’t always the case with science. It is often in the interest of researchers to not release their data, largely because they can then publish more papers using these datasets without others scooping them. Our goal is to let as many people use our data as possible. 

We also believe that this scientific revolution through big data is just the beginning. The initial round of research papers focused on basic measures of fishing effort. We can do much more. For instance, we are now working with two different labs to identify likely nefarious activity based on how vessels behave — with the idea that vessels that fish illegally move about the oceans in a different way from those that operate legally (for instance, they may turn off their AIS device more frequently). We are also partnering with labs that are attempting to measure the amount of fish caught and the type of fish — moving from measuring hours of fishing to tons of fish caught, which is much more useful for fisheries managers. And perhaps most excitingly, we are working with new global satellite radar technology to track all industrial fishing vessels in the world, not just the vessels that are (voluntarily) broadcasting their location. The same principles that allowed the first wave of research to be successful should, hopefully, apply to this second round.

Ultimately, GFW is founded on a simple idea: What happens in the world’s fisheries should not be a mystery.

The harvesting of seafood — the removal of a publicly owned resource from the world’s oceans — should be monitored and understood by the public. By making information about fisheries widely available, we can help identify and end the unsustainable and illicit activity plaguing many of the world’s fisheries.