33  Spatial Region Sources

Modified

February 20, 2025

33.1 Raster information

Raster Information can be found in pipe_static.spatial_measures_clustered_20230307

33.2 Regions

Regions are polygon layers that are used in the later stages of the pipeline to determine whether a given position (such as the location of a fishing event) is inside or outside the individual polygons in the layer. Example layers are WDPA marine protected areas, EEZs, RFMO regions, etc.

33.3 Description of Region Layers

Each set of region layers comes from a different source and is composed of a set of polygons that define a number of regions within that layer. Each layer has a unique ID field that identifies the real-world feature which may be represented by multiple polygons. For instance, the original EEZ layer has 323 polygon features which represent 281 distinct EEZs. See the child pages for details on where the layers come from, attribution and citation requirements and how to access the data.

33.3.1 Available Regions Layers

Pipeline bucket: pipe_regions_layers

  • [12 NM Boundary]
  • [EEZ Boundaries]
  • [FAO Areas]
  • [High Seas]
  • [High Seas Pockets]
  • [Oceans and Seas]
  • [Protected Seas]
  • [WDPA Marine MPAs]
  • [HOWTO: Do point in polygon test in bigquery using regions]
  • [HOWTO: Update a layer in pipe_regions]

33.4 Data Access

All these regions are managed via the pipe-regions tool. Regions are available in GCS and Bigquery with all attributes from the source shapefiles.

Each layer is normalized into a single zipped shapefile which is stored in GCS here gs://pipe-regions-layers-us-central1/

Layers are then converted to geojson and loaded into bigquery here: world-fishing-827.pipe_regions_layers

Layers are also processed to optimize for point in polygon testing using S2 cells in bigquery. The optimized tables are in world-fishing-827.pipe_regions_layers in tables with suffix **_S2_10** which indicates it is optimized for S2 cells at level 10.

33.5 Background

Previously regions were included in the Spatial Measures pipeline stage, along with other scalar measures like bathymetry depth and distance from shore. There were some problems with the old approach that have been fixed with the new implementation in pipe-regions.

  • In Spatial Measures regions were gridded to 1km resolution, so positions near the boundary were sometimes mis-classified as being inside or outside. This gridded approach was a compromise made for performance reasons. With the new pipe-regions implementation, every lat/lon position is tested against the full resolution polygon using ST_INTERSECTS, so it is very accurate for positions near the polygon boundary
  • The Spatial Measures step comes early in the pipeline, so this meant that updating the polygons used for computing regions was expensive since all downstream pipeline stages would need to be re-run. The WDPA is updated monthly, and we would like for these monthly updates to be automatically included in the pipeline. The new pipe-regions implementation comes later in the pipeline, and has enhanced support for automated updates of polygon layers, so it is easier to make changes and keep the pipeline output up to date
  • Mismatch between compute and display - the polygons that are displayed on the map are the true polygons, and this means that with the Spatial Measures approach, there was a mismatch between what was computed and what was displayed.

33.6 Legacy Regions

List of regions that were present in the old spatial_measures table. Need to figure out what to do with these.

Reference docs

legacy layer id from Spatial Mesaures New Regions S2 layer
mregion Not sure what this is. there are 312 unique id values
arg Argentina?
eez EEZ Boundaries
mparu I think this is a subset of the WDPA MPA layer - restricted use?
vme ???
kkp Indonesia
ocean oceans from marine regions. Looks like this is just the 5 main oceans, but marine regions has some more detailed layers - World Seas IHO with 101 features and GOAS with 10 features. Does anyone even use this?
hsp high seas pockets (marine regions?). I think this was a custom layer we created.
other does not seem to be used
fao FAO Areas
mpant I think this is a subset of the WDPA mpa layer. MPA No Take.
ames ??? 207 values - ARC_1..11, ETTP_1..21, MED 1..17, NP_1..20, NWA_1..7, SEA_1..45, SIO_1..39, and so on…
rfmo RFMOs
major_fao FAO Areas
WDPA Marine MPAs

Checking to see what RFMOs are being used.

SELECT

SELECT
distinct rfmo

FROM `world-fishing-827.pipe_static.spatial_measures_20201105`

CROSS JOIN UNNEST(regions.rfmo) as rfmo

LIMIT 1000

This gives a list of 42 RFMO labels. All of these have been incorporated into the new RFMO layer RFMOs

Legacy RFMO identifier

ACAP

APFIC

BOBP-IGO

CACFISH

CCAMLR

CCBSP

CCSBT

COREP

CPPS

CRFM

CTMFM

EIFAAC

FCWC

FFA

GFCM

IATTC

ICCAT

ICES

IOTC

IPHC

IWC

LTA

LVFO

MRC

NACA

NAFO

NAMMCO

NASCO

NEAFC

NPAFC

NPFC

OSPESCA

PERSGA

PICES

RECOFI

SEAFDEC

SIOFA

SPC

SPRFMO

SRFC

SWIOFC

WCPFC

33.7 Design Issues

https://github.com/GlobalFishingWatch/pipe-regions/blob/master/DESIGN.md