Mapping nighttime fishing in Southeast Asia

Nighttime fishing in Southeast Asia, where the boats are small, relies on the use of extremely bright lights to attract fish and the boats are not obliged to report GPS locations. However, they can be detected and tracked from space using a high-sensitivity telescope. In this project led by the University of Colorado, Boulder, an innovative method was developed to robustly detect fishing boats at night using visible and infrared images from the Visible Infrared Imaging Radiometer Suite (VIIRS) sensor onboard the Suomi National Polar-orbiting Partnership (SNPP) satellite.

This provided government agencies, the scientific community and environmental groups with maps and time series summarizing long term nighttime fishing activity. These maps and statistics were the scientific basis for SDG 14.4 to effectively regulate harvesting and to end overfishing. With this data, it became possible to estimate fishing environmental effects and trends and assess the efficacy of some of the regulatory measures which restricted fishing by date, boat location, size, and light power.

A pilot version of the VIIRS Boat Detection Database (VBD DB) used Postgres SQL with GIS extensions as a database management system. The total number of boat detections in the database was around 5,000,000. Tools were also developed for visualization and statistical analysis of this extremely dense point data set with a projected size of 50,000,000 boat detections in 2012-2017 with global coverage. This analysis toolkit was web-based so users could easily share the visualizations and peer-review the analysis algorithms. This project addressed SDG 14 - Life below water.  A retrospective VIIRS Boat Detections Dataset was created for years 2012-2017 and added to the database. Using that retrospective database, a new index could measure the efficiency of the fishery closures. 

In June 2018, the web visualization toolkit was extended to observe and statistically prove changes in the fishing activity pattern. Google Earth Engine was chosen for web-based GIS and Visualization toolkit for two reasons: 

  1. It has a rich ready-to-use JavaScript library of GIS functions and,

  2. The possibility to share data with other research projects on the same platform, for example, the Global Fishing Watch project1. 

The resulting visualization service, named the Google Earth Engine Assessment Tool, was presented at the Global Fishing Watch scientific conference in June 2018.

Based on the VIIRS Boat Detections Dataset and GIS maps of Marine Protected Areas (MPAs), an Illegal Fishing Alert System was tested and developed. It utilizes e-mail, SMS and Telegram messengers to disseminate alerts of illegal fishing boat detections from space to local law enforcement units in near-real-time. 

Disclaimer: As many of the links are external, some links may not work.