Dataset Viewer
Auto-converted to Parquet Duplicate
Search is not available for this dataset
The dataset viewer is not available for this split.
Server error while post-processing the rows. This occured on row 42. Please report the issue.
Error code:   RowsPostProcessingError

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

TORRENT — portal asset backend

Rendered images and download packages for the TORRENT flash-flood benchmark. This repository is not a standalone dataset: it is the static asset backend that the TORRENT portal Space loads images and zip bundles from over the Hub CDN.

Everything here is derived from the frozen Zenodo v1.0 archive. If a number in a figure disagrees with the archive, the archive wins.

Layout

<ff_episode_id>/<gage_id>/figures/figure5.png     4-panel benchmark overview
<ff_episode_id>/<gage_id>/overlays/mrms_total.png MRMS event-total raster overlay (RGBA)
downloads/<ff_episode_id>__<gage_id>.zip          one testbed, entries under '<gage_id>/'
downloads/torrent_primary21_v1.0.zip              all 21, entries under '<ep>/<gage>/'

gage_id keeps its leading zeros (e.g. 03165000); testbed_id is <ff_episode_id>__<gage_id>.

Coverage is the 21 primary-benchmark testbeds listed in data/L3/primary_benchmark.csv of the Zenodo archive.

figure5.png panels

panel content
(a) DEM hillshade + NHD stream network + watershed + NCEI reports + temporally matched HWMs + gauge
(b) Observed USGS discharge + MRMS basin-mean precipitation (inverted axis), peak markers
(c) MRMS event-total precipitation, watershed-clipped, on hillshade
(d) CREST temporal-max unit streamflow sampled onto NHD flowlines

mrms_total.png

Transparent RGBA, nearest-neighbour upscaled 12x, MRMS-QPE colour ramp, colour range = 2nd–98th percentile of the in-grid event total. Geographic bounds and the colour range for each testbed live in the Space's data/overlay_bounds.json as [south, west, north, east] and mrms_range.

Missing metrics on three testbeds

FF_2024_09_NC_ep007/03165000, FF_2025_05_MD_ep001/01596500 and FF_2025_06_TX_ep002/08185065 entered the benchmark through NLDI basin recovery. Their pre_event_baseline, time_to_peak_hr and flash_flood_class are genuinely absent in the frozen v1.0 tables. Their figures therefore print n/a in those slots and omit the pre-event baseline reference line — no value has been substituted or estimated.

Citation

Cite the Zenodo archive, not this repository: https://doi.org/10.5281/zenodo.22118051.

Figures and overlays are released CC-BY-4.0. The underlying USGS, NOAA/NCEI and MRMS/FLASH source data are US Government works in the public domain.

Downloads last month
28

Spaces using skyan1002/flash-flood-benchmark-figures 2