# Data Sources — Petrel Flow Accumulation Layer v0.3 All input data used to build the v0.3 vintage, with citations and licensing notes. Every source below is **attribution-only** (CC BY 4.0 or public domain): no NonCommercial term, no share-alike, no EULA flow-down, nothing that reaches a derived raster. That is what lets Petrel own this Layer and both give it away and embed it in commercial hazard models — see [`model_card.md`](model_card.md). v0.3 is a **deterministic derivation**, not a trained model: there is no labelled training set, only the inputs and the published methods below. ## 1. GEDTM30 v1.2 — elevation (the terrain input) Global Ensemble Digital Terrain Model, 30 m, **bare earth**. Buildings and canopy are removed, which matters here because neither dams a river. - **Citation:** Ho, Y.-F., Hengl, T., et al. (2025). *GEDTM30: global ensemble digital terrain model at 30 m and derived multiscale terrain variables.* PeerJ 13:e19673. https://doi.org/10.7717/peerj.19673 · concept DOI 10.5281/zenodo.14900180 · v1.2 record https://zenodo.org/records/18887460 - **License:** **CC BY 4.0**, verified at the Zenodo API (`license.id` = `cc-by-4.0`) for v0, v1 and v1.2, and in the README §License verbatim: *"This dataset is released under fully open license CC-BY 4.0."* - **Attribution (required, reproduce downstream):** *"GEDTM30 v1.2 © OpenGeoHub / Ho et al. 2025, CC BY 4.0"*, plus the upstream **Copernicus WorldDEM production and liability notices** and **JAXA AW3D30** attribution that GEDTM30 itself inherits. - **Vertical datum:** EGM2008 (EPSG:4326+3855). - **Used as:** the elevation grid fed to pysheds (fill → flow direction → accumulation), the slope term of TWI, and the depression-depth scan that identifies closed basins. Read at 9 arcsec (~278 m) for the global Layer. - **Lineage note, recorded for honesty:** GEDTM30's own production code uses MERIT DEM (CC BY-NC / ODbL) as a *training-point QC filter* only — it is commented out of the covariate list and no MERIT value flows into the raster. FABDEM (CC BY-NC) appears only in the authors' validation notebooks, not as a covariate; the withdrawn 2023 EDTM v1 is a different release. This is the authors' licence chain, not ours, and their release is CC BY 4.0. ## 2. HydroLAKES v1.0 — reservoir polygons and pour points - **Citation:** Messager, M. L., Lehner, B., Grill, G., Nedeva, I., Schmitt, O. (2016). *Estimating the volume and age of water stored in global lakes using a geo-statistical approach.* Nature Communications 7: 13603. https://doi.org/10.1038/ncomms13603 - **License:** **CC BY 4.0**, verified at the product page 2026-09-17: *"The HydroLAKES database is licensed under a Creative Commons Attribution (CC-BY) 4.0 International License."* Attribution only; no NC, no share-alike. - **Used as:** the polygon and the expert outlet (pour point) for each of the eight Volga/Kama reservoirs treated as flow nodes, plus `Wshd_area` as an independent check on what each should carry. See [`methodology.md`](methodology.md) → *Reservoirs as flow nodes*. - **Not HydroSHEDS core / HydroRIVERS.** HydroLAKES carries its own CC BY 4.0 licence on its own product page; that is the ruling relied on here. The HydroSHEDS v1 terms discussed below do not apply to it. ## 3. Natural Earth 1:10m rivers and lake centerlines - **Source:** https://www.naturalearthdata.com/downloads/10m-physical-vectors/10m-rivers-lake-centerlines/ - **License:** **public domain.** Natural Earth terms of use: *"no permission is needed to use Natural Earth. Crediting the authors is unnecessary."* Credited here anyway. - **Used as:** the mapped course a river takes **through** a reservoir, where flat water leaves the DEM no gradient to follow. For each reservoir node the exit corridor is traced along the centerline from the dam to the next reservoir or to open river. This shapes flow direction on those reaches, so it is a data input to v0.3, not merely a validation reference as in v0.2. ## 4. GSHHG v2.3.7 full-resolution shoreline (`gshhg_f`) Petrel's canonical land mask — Wessel & Smith's Global Self-consistent, Hierarchical, High-resolution Geography database at full resolution (~100 m). - **Citation:** Wessel, P., & Smith, W. H. F. (1996). A global, self-consistent, hierarchical, high-resolution shoreline database. *Journal of Geophysical Research*, 101(B4), 8741–8743. - **License:** LGPL (internal masking use only; not redistributed as data). - **Used as:** the coast clip applied to both rasters (`all_touched` — any cell touching real land keeps its value) and the land mask that keeps the flow network off the ocean. GSHHG L1 only, so Antarctica is excluded by standard. ## Why not HydroSHEDS / HydroRIVERS? The classical drainage layers (distance-to-drainage, flow accumulation, river network) usually come from HydroSHEDS / HydroRIVERS. Verified 2026-09-07 against the WWF technical documentation: - **HydroSHEDS v1 / HydroRIVERS: unusable, and not because of NonCommercial.** §7.1 explicitly permits commercial use. The blocker is §3, which assigns WWF ownership of *"all Intellectual Property Rights in the Licensed Materials and any modifications, extensions, or improvements ... however developed"* — that reaches a Petrel-derived raster directly. Plus an End-User EULA flow-down onto our customers and a ban on stand-alone redistribution (§2.1.2). "Our derivative is a new work" is exactly what §3 forecloses. - **HydroSHEDS v2 (RIV_STREAMS) is CC BY 4.0 and IS clean.** §4.1: *"All products of version 2 of HydroSHEDS are distributed under a Creative Commons Attribution 4.0 International License."* It carries full reach connectivity at 1 arcsec on a 1 km² threshold — but it is **Americas-only** as of TechDoc v2.0.0 (July 2026), so it cannot be a global spine. Usable as a validation reference where it covers. - **MERIT Hydro remains excluded** (CC BY-NC 4.0 or ODbL — NC fails, share-alike fails). This Layer re-derives the same quantities clean-room from attribution-only inputs, which remains the only route that is simultaneously global, licence-clean and Petrel-owned. ## Method references (no data, published algorithms) ### pysheds - **Citation:** Bartos, M. (2020). *pysheds: simple and fast watershed delineation in Python.* https://github.com/mdbartos/pysheds · https://doi.org/10.5281/zenodo.3822494 - **License:** GNU GPL-3.0 — used as a **processing tool** (imported at build time); it processes our data and is not redistributed or linked into the shipped raster, so its terms do not attach to the output data product. - **Used as:** the depression-handling + D8 flow-direction + accumulation engine. The Euclidean distance transform, the connected-component labelling used to fill reservoir bodies, and the slope/gradient term of TWI use SciPy / NumPy (BSD-licensed numerical libraries) — again tools, not data sources. ## Reproducibility The exact source paths, method knobs (resolution, halo width, channel threshold, caps, reservoir node table) and promote configuration for v0.3 are versioned in `configs/flow_accumulation_volga.yaml` in the Petrel ML repository. That config is the source of truth — if anything here disagrees with it, the config wins.