VIIRS Boat Detection

"The light from our glass coils produced magical effects at times, lingering on the wrinkled roughness of some natural arch, or some overhang suspended like a chandelier, which our lamps flecked with fiery sparks."

Jules Verne - 20,000 Leagues Under the Sea - 1872


SNPP is the Suomi National Polar Partnership satellite flown by NASA and NOAA. It is the next generation polar orbiting satellite, collecting both daytime and nighttime data worldwide each day. The primary imager on SNPP is the Visible Infrared Imaging Radiometer Suite (VIIRS). At night the VIIRS day / night band (DNB) collects low light imaging data in the visible to enable detection of moonlit clouds. In addition to clouds, the DNB detects lights sources present at the earth's surface. This includes the detection of boats, especially those using light to attract catch. Earth Observation Group (EOG) has worked since 2014 on algorithms to report the locations of boat detected based on lights with low temporal latency. The service started in S.E. Asia and has now been expanded to global.

Data Format

VBD data are available in multiple formats and different temporal aggregations. The data is available for individual country EEZs and global*.

* Only selected countries. Area impacted by SAA is redacted in global coverage products.

Nightly Data

Nightly data is available in CSV and KMZ format. For KMZ files, it is recommended to use Google Earth to browse the content. 


CSV readme file is available HERE.

Quality Flags (QF)

Each VBD detection is given a quality flag to mark its quality as a vessel detection, and possible types if it is not deems not a lit vessel.

KMZ Placemark Quality Flag Definition
1 Boat
2 Weak Detection
3 Blurry Detection
4 Gas Flare
7 Glow
8 Recurring Light
10 Weak and Blurry
11 Platform

Monthly Summary Grid

Data are aggregated monthly for country level and global products in GeoTIFF format in 15 arc second resolution. Each set of summary grid zip file contains 4 GeoTIFF files.

Below is a example file set for Philippines during August 2020.

File naming convention explained:
VBD_[satellite]_[start date]-[end date]_[region]_[configuration]_[version]_c[creation time].[flavor].tif

Satellite Name of the satellite collected the imageries.
npp -> Suomi NPP
j01 -> NOAA-20 (JPSS-1)
Start/End Date Start and end date of the aggregation in format YYYYMMDD.
Region Region of the product. For country level products, 3 digit code from ISO-3166 is used. For global it is global-saa.
Configuration The quality flag filter used in the aggregation.
Version Version of the VBD algorithm software.
Creation Time Time of file creation, in format YYYMMDDhh.
Flavor Flavor of the product.
cvg -> Count of satellite overpass coverage.
avg_rade9 -> Average radiance of the detection (nW/cm2/sr).
n_detect -> Count of VBD detections.
pct_detect -> Percentage of VBD detections (n_detect/cvg).

Annual Summary Grid

The summary grid is also available in annual aggregation.


A handy navigator is designed to facilitate users to find the desired VBD product other than direct access to the repository. VBD product is produced in near real-time (NRT) latency. Typically the record will be available in 6 hours after satellite overpass. When a complete UTC day is collected, the file will be moved to another directory for archive. Output from NRT data stream are reserved for paid subscribers. Please contact EOG for details if interested. EOG also reprocesses VBD product every month with the output placed in a separate directory for open access. See below for a illustrated flow of file availability.

File Naming

The CSV and KMZ file names looks like below.
(CSV) VBD_npp_d20120401_idn_noaa_ops_v23.csv
(KMZ) VBD_npp_d20120401_idn_noaa_ops_v23.kmz

Fields are broken down as below.

Satellite Name of the satellite collected the imageries.
npp -> Suomi NPP
j01 -> NOAA-20 (JPSS-1) *
* VBD is not yet processing imagery from NOAA-20.
Date UTC date of the file in format YYYYMMDD.
Region Region of the product. For country level products, 3 digit code from ISO-3166 is used. For global it is global-saa for SAA redacted version, and global for non-redacted version.
Version Version of the VBD algorithm software.

Files in Navigator:
Files shown in Navigator are given labels regarding their state of process.
[CURRENT] -> Data is still being updated in near real-time. Restricted access.
[COMPLETE] -> Data for a complete UTC day. Restricted access.
[FINAL] -> Reprocessed data. Open access.

File in Repository:
Files in repository are accessible with link composed as below.[region]/[state]/[file]

region Region of the product. For country level products, 3 digit code from ISO-3166 is used. For global it is global-saa for SAA redacted version, and global for non-redacted version.
state current -> Data is still being updated in near real-time.
nrt (or complete) -> Data for a complete UTC day.
fnl (or final) -> Reprocessed data. Open access.
file File name.


Users can choose to find and download data with navigator interface or directly go to repository.

South Atlantic Anomaly (SAA)

The global VBD data is redacted over South America and its nearby waters. This is because the anomalous spike of radiance caused by the high energy particles from the space striking the VIIRS DNB sensor causing false VBD detection. Such artifact is often observed over the SAA region as shown in the map below. EOG had redacted the most severely affected area to avoid misleading users, while working on advanced algorithm to improve the result over SAA region.

Read more about SAA HERE.

Temporal Loops

Check out the movie loops EOG made for different regions of interest showcasing changes of VBD over time.

Open VBD Temporal Loops


When using the data please credit the product generation to the Earth Observation Group, Payne Institute for Public Policy, with proper citations as below.

  • Hsu, F.C., Elvidge, C.D., Baugh, K., Zhizhin, M., Ghosh, T., Kroodsma, D., Susanto, A., Budy, W., Riyanto, M., Nurzeha, R. and Sudarja, Y., 2019. Cross-matching VIIRS boat detections with vessel monitoring system tracks in Indonesia. Remote Sensing11(9), p.995.
  • Elvidge, C.D., Ghosh, T., Baugh, K., Zhizhin, M., Hsu, F.C., Katada, N.S., Penalosa, W. and Hung, B.Q., 2018. Rating the effectiveness of fishery closures with visible infrared imaging radiometer suite boat detection data. Frontiers in Marine Science5, p.132.
  • Elvidge, C.D., Zhizhin, M., Baugh, K. and Hsu, F.C., 2015. Automatic boat identification system for VIIRS low light imaging data. Remote sensing7(3), pp.3020-3036.


Global Fishing Watch