Skip to contents

Downloading BC Spatial Data

Source: vignettes/data-sources.Rmd
data-sources.Rmd

Overview

rfp_source() downloads BC spatial data layers into a GeoPackage, clipped to one or more watershed groups. Six source types are available; the authoritative list is rfp_manifest_types(), of which every value but frozen is refreshable:

Type Source What you get
bcdata BC Data Catalogue PSCIS crossings, Indian reserves, BC regions, and layers listed in inst/lookups/rfp_source_bcdata.txt
fwa Hillcrest Geo FWA feature service Freshwater Atlas named streams, watersheds
aws NewGraph S3 bucket bcfishpass crossings/streams, transport, cadastre, lateral habitat raster, model parameters
osm OpenStreetMap via Overpass trails and other layers with no provincial source, per inst/lookups/rfp_source_osm.csv
url any address ogr2ogr can open federal open data, a provincial WFS, per inst/lookups/rfp_source_url.csv
stac a STAC catalogue N items matched by watershed group and item filter, one asset each, merged into a single layer, per inst/lookups/rfp_source_stac.csv

The three CSV lookups (osm, url, stac) are defaults. rfp_source(lookups = ) takes a caller’s own CSV per tier, merged row-wise over the shipped rows and keyed on source_layer — so swapping a STAC collection, adding an epoch or retuning an Overpass selector is a commit in the repo that owns the config, not an rfp release and a reinstall on every host. rfp_source_lookup() returns the merged result, with a lookup_source column naming the file each row came from.

Basic usage

Download the classic three source types for a watershed group:

library(rfp)

rfp_source("bcdata", c("ADMS"))
rfp_source("fwa", c("ADMS"))
rfp_source("aws", c("ADMS"))

Each call appends layers to background_layers.gpkg in the current working directory. The aws type also produces habitat_lateral.tif.

Multiple watershed groups

Pass a character vector with multiple codes:

rfp_source("bcdata", c("BULK", "KLUM"))

Custom output path 

rfp_source("bcdata", c("ELKR"), path_gpkg = "my_project.gpkg")

Configuring which layers to download

Layer lists are text files in inst/lookups/. Each line is a layer name. Lines starting with # are disabled:

# Active layers (uncommented)
whse_fish.pscis_assessment_svw
reg_legal_and_admin_boundaries.qsoi_bc_regions

# Disabled layers (commented out)
#whse_basemapping.bcgs_20k_grid
#whse_forest_vegetation.bec_biogeoclimatic_poly

To customise, copy the text file from inst/lookups/ to your project directory and edit it.

How it works

Under the hood, rfp_source():

  1. Formats watershed group codes for the shell (e.g., c("BULK", "KLUM") becomes "'BULK', 'KLUM'")
  2. Locates the shell script and layer list in the installed package via system.file()
  3. Sets environment variables (RFP_GPKG, RFP_SOURCES, RFP_TARGETS, …) for the shell script
  4. Calls the script via system2("bash", ...)

The shell scripts call ogr2ogr, wget and unzip to download, clip and load layers into the GeoPackage. The aws pass additionally uses rio for the lateral-habitat raster; rfp resolves that binary and passes it in as RFP_RIO, along with RFP_PROJ_LIB scoped to that one call. Nothing is placed on PATH and PROJ_LIB is not set globally, so ogr2ogr keeps using its own PROJ data. Run rfp_check_deps() to see what resolves on your machine.

Single-layer download (pure R)

For downloading a single BC Data Catalogue layer without Python CLI tools, use spacehakr::spk_source_bcdata():

spacehakr::spk_source_bcdata(
  bcdata_record_id = "whse_fish.pscis_assessment_svw",
  path_gpkg = "background_layers.gpkg",
  mask = my_aoi_polygon
)

This uses the bcdata R package directly — no shell scripts, no Python.

The function used to live here as rfp_source_bcdata(). It moved to spacehakr because fetching a published dataset is not specific to this package, and sourcing public data is what makes a report’s inputs reproducible for a reader who does not have this package installed. rfp_source_bcdata() still works and forwards there, with a deprecation warning.