While many of us have heard how big data and the Internet of Things is touching every area of business, innovative companies are starting to use big data analytics to solve environmental problems as well. A data science company in the Washington DC area called Analyze is applying their deep behavioral analytic tradecraft to a long-standing global problem: illegal fishing.
Illegal, unregulated and unreported (IUU) fishing is a significant economic and environmental challenge for countries around the world. It is estimated that up to 40% of fishing catch in certain parts of the world is unlawful or unregulated, resulting in approximately $10B to $20B in economic losses to legitimate fisheries and significantly depleting international food stocks.
Analyze, a company of data scientists specialized in high-end behavioral and predicative data models, have developed a novel, reliable method for characterizing fishing behaviors among ships on the high seas based on geospatial position information reported through the Automated Identification System (AIS). The method, when integrated into a broader system of vessel monitoring, behavior detection and enforcement, has the potential to significantly improve interdiction of illegal fishing on the high seas. Even without the integrated system the method offers scientists new ways to measure fishing capture yields and monitor the change of fishing behavior over time and inform cause and effects of fishing on fish populations and their ecosystems.
By observing long-term vessel movement patterns of all the ships on the ocean and employing a feature reduction and behavior classification-based machine learning approach, the method, which Analyze implemented through a python software program termed Mercury, makes reliable determinations as to whether a ship is fishing as distinct from other maritime activities such as cargo transiting, passenger service, tug and rescue activity and maritime law enforcement. Their approach relies on the semi-voluntary reporting of position information via the AIS system, which is received and processed by Analyze via low earth orbit satellite. It does not depend, however, on a vessel’s self-reporting of fishing (Navsat 7) or IMO classification as a fishing vessel. Only geometric patterns of motion are used to make a behavior determination.
The automated identification system (AIS) is a VHF-frequency automated aid to navigation require for vessels over 90 tons and used many smaller vessels. Originally conceptualized for collision avoidance, AIS is used for a number of applications related to monitoring of ship movements. The system was originally designed to be operated via shore and vessel-based receivers and an extensive network of coastal sensors has been developed for this purpose. The ease of reception of VHF frequencies by low earth orbit satellites, however, has led to the introduction of space-based AIS sensing. Space-based sensors have several advantages over land-based AIS receivers, most significantly, the ability to detect AIS messages beyond thirty miles from shore based sensors.
Ed Lorenzini, President of Analyze, commented, “We are excited about applying our expertise to help combat illegal fishing. This project gave us the opportunity to bring a fresh perspective, be innovative, and ultimately let our algorithm speak for itself. We appreciate working on these types of difficult problems because it’s difficult for first movers to overshadow our data science techniques and technologies with less sophisticated models.”
Mr. Lorenzini explained that a year ago they were presented with a challenge by none other then Google. Google needed all of the ships on the ocean analyzed so they could showcase data on Google Oceans, Big Query, and Big Store—Google’s platform for Big Data—at the 2013 Google IO Conference.
In June 2013, shortly after Google showcased its work, Analyze hosted the first IUU Fishing Roundtable. With attendees from the United States, United Kingdom, and Israel, the Roundtable brought together the most influential players in the campaign against IUU fishing, including the National Oceanic and Atmospheric Administration (NOAA), Pew Charitable Trusts, Google, SpaceQuest, Greenline Systems, IHS Fairplay, Windward Maritime Solutions, OrbComm, and SkyTruth. You can read more about the meeting by downloading a report from Analyze’s web site.
After the Roundtable, Analyze took the raw sensor data collected from space over a four year period by SpaceQuest and found that with more than 100,000 ships broadcasting their location at any one time, every 6 seconds, yielding as many as 18 million data records per day it could use an algorithm to classify vessels based on their movement patterns, despite the semi voluntary nature and sporadic use of AIS by some ship captains.
NOAA led Analyze to the Pew Charitable Trusts and another non-profit called SkyTruth to explore technologies that can address maritime surveillance and monitoring needs for ocean conservation purposes. SkyTruth leaned on Analyze’s expertise to develop and demonstrate the automated classification of vessel behavior from AIS tracking data.
The story has now come full circle as Google is preparing to host AIS data and Analyze’s algorithm in its cloud and make Big Query available to support an open system on the web that will enable global visibility into fishing on the open ocean.