|
Download README.md from gtmenichino/staf-data: direct link, hf CLI and curl.
- Browser
- Download file 2.01 kB
-
https://huggingface.co/datasets/gtmenichino/staf-data/resolve/main/README.md
- Command line
-
hf download hf://datasets/gtmenichino/staf-data/README.md
-
curl -L -o README.md https://huggingface.co/datasets/gtmenichino/staf-data/resolve/main/README.md
2.01 kB
| pretty_name: STAF data (browser mirror) | |
| tags: | |
| - hydrology | |
| - streams | |
| - nhdplus | |
| - geospatial | |
| # STAF data: the browser mirror | |
| The data the static STAF web app (EASI, SFARI and DEEP; the Stream Tiered Assessment Framework, | |
| https://usace-wrises.github.io/staf/) reads by HTTP range. Every file mirrors an asset of a GitHub | |
| release of https://github.com/USACE-WRISES/staf, which stays the source: | |
| - `bundle/`: the STAF data bundle (`staf-data-current`), unzipped: NHDPlus HR flowlines, catchments | |
| and their shared borders, precomputed per-catchment values, the NHDPlus V2 crosswalk with | |
| StreamCat, and national point tables (WQP, NID, NWIS, NAS), for the lower 48. | |
| - `xs/`: the national cross-section grid (`staf-xs-current`, published region by region): every | |
| section's scoring numbers, what drawing or pulling a section needs, one profile per segment. | |
| - `points/`: the bundle's national point tables (dams, stream gages, water-quality stations and | |
| results) sorted by place in small row groups, so a lookup near a site reads a few kilobytes. | |
| - `library/`: the STAF assessment library (`library`): DEEP's regional assessments, calculators and | |
| EASI's method packages. | |
| - `engine/`: the data files the STAF site engine ships with: the USGS base-flow index grid (Wolock | |
| 2003), the EPA Level III ecoregions and the physiographic divisions. | |
| The files the app reads a few rows of at a time (each region's stream lines, StreamCat rows, V2 | |
| attributes and point extras, and the cross-section files) are written again with the same rows and values | |
| in small pages with a page index, so a read takes kilobytes; `topo/` holds each region's network topology | |
| on its own. JSON and GeoJSON files are stored gzipped (`.gz`). | |
| Files are stored under content-hashed names and never change in place; `web-manifest.json` (updated | |
| last) maps each logical name to its file, size and sha256. Built from public federal datasets (USGS, | |
| EPA, USACE, NOAA, USDA, FWS); see the STAF repository for sources and methods. | |