Skip to main content
All CollectionsFisheries analysis
How it works: Fishing classifier model
How it works: Fishing classifier model
Andy Hovey avatar
Written by Andy Hovey
Updated over a week ago

How fishing classifications are reflected in the app

The fishing classifier model is applied to the track of every vessel with an AIS vessel class of fishing or unknown. Portions of the track where the activity is determined by the model to be fishing, are coloured pink. Start and stop times for each continuous fishing event are listed in the track history of fishing vessels.

How our fishing classifier works

Our fishing classifier is based on a version of Global Fishing Watch’s (GFW) model described in Kroodsma et. al. 2018¹.

The model uses a machine learning algorithm known as a convolutional neural network to classify a time series of movement related features for each vessel. The features include information such as vessel speed and course, local time and month, distance from the nearest shore, and time between AIS messages.

When developing the model, GFW worked with fishing experts to classify fishing activity². They evaluated 247,000 hours of AIS tracks from 624 vessels, capturing a range of fishing methods. The model was trained to distinguish between fishing and non-fishing activity using 503 of these vessels’ tracks. The tracks from the remaining 121 vessels were used to evaluate the model’s accuracy. Regardless of the type of fishing being undertaken (e.g. longlining, trawling, purse seine), the model was able to correctly classify a vessel as fishing or not fishing more than 90% of the time.

The Starboard fishing classifier has an additional post-processing step—for any two hour period where fishing is detected we classify the entire two hour period as fishing. This is done to merge any short fishing events that occur near each other into one fishing event, providing a more realistic representation of fishing behaviour. However, it will also have the effect of expanding the fishing event past the start and end times.

The fishing classifier is processed hourly and potential fishing events that have occurred as recently as four and a half hours ago can be classified.

Frequently asked questions

Why is this vessel that is clearly fishing classified as not fishing? And likewise, why is a vessel that is clearly not fishing classified as fishing?

Vessels self-report their vessel type in AIS, such as cargo, fishing or pleasure craft. If the vessel hasn’t self-reported this information their AIS vessel type is reported as unknown. Starboard only shows fishing events for vessels whose AIS-reported vessel type is fishing or unknown, or we have determined it to be a fishing vessel based on the vessel’s track and behaviour. This means that occasionally vessels with an incorrect AIS-reported vessel type may be misclassified as fishing or not-fishing. As we learn about these we update the AIS-reported vessel type in Starboard to reflect the actual vessel type.

Why are your results different from Global Fishing Watch’s results?

This is likely due to differences in the AIS datasets used. We use Spire Real-Time AIS (previously ExactEarth) and S&P Global Market Intelligence as our AIS suppliers and GFW uses Orbcomm and Spire.

References

  1. Kroodsma, David A., Juan Mayorga, Timothy Hochberg, Nathan A. Miller, Kristina Boerder, Francesco Ferretti, Alex Wilson, et al. “Tracking the Global Footprint of Fisheries.” Science 359, no. 6378 (February 23, 2018): 904–8. https://doi.org/10.1126/science.aao5646.

  2. Souza, Erico N. de, Kristina Boerder, Stan Matwin, and Boris Worm. “Improving Fishing Pattern Detection from Satellite AIS Using Data Mining and Machine Learning.” Edited by Athanassios C. Tsikliras. PLOS ONE 11, no. 7 (July 1, 2016): e0158248. https://doi.org/10.1371/journal.pone.0158248.

Did this answer your question?