Import Bio-ORACLE v3.0 ocean environmental layers (present + future climate)

[boettiger-lab-llm-agent]

Dataset

Bio-ORACLE v3.0 — Global marine environmental rasters at 0.05° (~5.5 km) resolution, present-day baselines (2000–2020) and CMIP6 future projections (2020–2100) under 6 SSP scenarios.

• Publication: Assis et al. (2024), Global Ecology and Biogeography — https://doi.org/10.1111/geb.13813
• Data server (ERDDAP): https://erddap.bio-oracle.org/erddap/
• Download page: https://www.bio-oracle.org/downloads-to-email.php
• R package: https://github.com/bio-oracle/biooracler
• Python package: https://github.com/bio-oracle/pyo_oracle
• License: GNU GPL

Variables (19 + terrain)

Group	Variables
Temperature	Sea water temperature (thetao), air temperature (tas)
Chemistry	Salinity (so), pH (ph), dissolved O₂ (o2), nitrate (no3), phosphate (po4), silicate (si), dissolved iron (dfe)
Productivity	Chlorophyll (chl), phytoplankton carbon (phyc), PAR (par), attenuation coeff (kdpar)
Circulation	Sea water speed (sws), direction (swd), mixed layer depth (mlotst)
Ice/Cloud	Sea ice thickness (sithick), sea ice cover (siconc), cloud cover (clt)
Terrain	Bathymetry (min/mean/max), slope, aspect, topographic position index, terrain ruggedness, coastline, landmass

Each variable has 6 statistics: max, mean, min, ltmax, ltmin, range.

Depth levels: surface (depthSurf), benthic (depthMax, depthMean, depthMin).

Temporal coverage:

• Present: 2000–2010, 2010–2020 baselines
• Future: decadal steps 2020–2100 under SSP1-1.9, SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP4-6.0, SSP5-8.5

Total datasets on ERDDAP server: ~356 (variable × depth × time period × scenario combinations).

Technical specs

• Resolution: 0.05° (~5.5 km), global 7200 × 3600 grid
• CRS: WGS84 (EPSG:4326)
• Native format: NetCDF-3, also served as GeoTIFF directly from ERDDAP
• No authentication required — ERDDAP REST API is fully curl/wget-friendly:

  # Full global GeoTIFF for mean sea surface temperature, 2000–2020 baseline
  curl -O "https://erddap.bio-oracle.org/erddap/griddap/thetao_baseline_2000_2020_depthsurf.geotif?thetao_mean%5B(2000-01-01T00:00:00Z)%5D%5B(-89.975):1:(89.975)%5D%5B(-179.975):1:(179.975)%5D"

  Full dataset list: https://erddap.bio-oracle.org/erddap/info/index.json?page=1&itemsPerPage=2000

Proposed processing

Scope question

With 356 datasets this is large. Likely approach is to be selective — prioritize:

1. Present-day baselines (2000–2020) for the most ecologically informative variables: thetao, so, o2, ph, chl, no3, bathymetry
2. Surface + benthic depth variants
3. Future scenarios for a key variable (e.g., temperature) under SSP1-2.6 and SSP5-8.5

Workflow

Each layer is a single-band global GeoTIFF → standard raster-workflow:

cng-datasets raster-workflow \
  --dataset bio-oracle/thetao-mean-surface \
  --source-url <erddap-geotif-url> \
  --bucket public-bio-oracle \
  --h3-resolution 8 \
  --value-column thetao_mean \
  --output-dir catalog/bio-oracle/k8s/thetao-mean-surface

Bucket: public-bio-oracle (new bucket)

Key use case

These layers are the canonical covariates for marine species distribution modeling (SDM). Combined with OBIS/GBIF occurrences (already in the catalog), they enable habitat suitability analysis for marine species directly in DuckDB via H3 joins.

Related issues / datasets

• GFW Fishing Effort (Import GFW Annual Fishing Effort (2012–2024) as H3 hex time series #52) — fishing pressure layer, joinable by H3
• Halpern Cumulative Human Impacts (not yet filed) — cumulative stressor layer
• GBIF occurrences (public-gbif) — species occurrences for SDM

[boettiger-lab-llm-agent]

Processing note: When hexing this dataset, include parent resolutions up to h5 (e.g. --parent-resolutions "5,6,7,8,9,0" or whatever subset applies). Parent resolutions are cheap to compute alongside the base resolution and significantly speed up spatial joins at coarser scales. For lower-resolution datasets (e.g. base res h6), still include h5 as a parent. Minimum: always include h5 and h0.