The Dataset Viewer has been disabled on this dataset.

D3P-JP-Q: Spatially distributed daily river discharge simulation over Japan

Version 1.0, simulated with the D3P-JP dataset, version 1.0.

This repository holds the most recent version of D3P-JP-Q and is the fastest way to download it. The snapshot taken at the time of publication of the paper is archived permanently on Zenodo: https://doi.org/10.5281/zenodo.24212556.

Simulated daily river discharge at the outlet of each of the 8,893 catchments of the D3P-JP river network, which covers Japan with catchments of about 40 km², from 2009-01-01 to 2020-12-30. The simulation is described in

Hascoet, T., Pellet, V., Oishi, S., & Miyoshi, T. Leveraging Japan's streamflow records for end-to-end learning of hydrological simulations at national scale.

The dataset it was trained on (river network, forcing, observed discharge) is distributed separately as D3P-JP (https://doi.org/10.5281/zenodo.24212554 and https://huggingface.co/datasets/tristhas/D3P-JP). Code and trained model: https://github.com/TristHas/jpdh.

Files

File Size Content
d3p-jp-q_simulation.tar 0.3 GB D3P-JP-Q.zarr, the simulation (Zarr format 3), and checksums/simulation.json, the SHA-256 of its files
catchment_outlets.csv the outlet of every catchment: location, drainage area, downstream catchment
MANIFEST.json sizes and SHA-256 checksums of the two files above

Download the two files and extract the archive:

wget https://huggingface.co/datasets/tristhas/D3P-JP-Q/resolve/main/d3p-jp-q_simulation.tar
wget https://huggingface.co/datasets/tristhas/D3P-JP-Q/resolve/main/catchment_outlets.csv
tar -xf d3p-jp-q_simulation.tar
import pandas as pd, xarray as xr
sim = xr.open_zarr("D3P-JP-Q.zarr")
outlets = pd.read_csv("catchment_outlets.csv", index_col="node_id")
node = outlets[outlets.drainage_area_km2 > 1000].index[0]         # pick a catchment
q = sim["discharge"].sel(spatial=node).to_series()                # m3/s, daily

D3P-JP-Q.zarr

Dimensions (time, spatial): 4,382 days and 8,893 catchments, single precision, no missing value.

Variable Content Units
discharge simulated discharge at the outlet of the catchment: the runoff of all its upstream catchments, routed along the river network m³ s⁻¹
runoff simulated runoff of the catchment's own area, before routing mm day⁻¹
spatial identifier of the catchment (node_id), shared with catchment_outlets.csv and with every file of D3P-JP –
lon, lat location of the outlet (WGS84), as in catchment_outlets.csv degrees
time 12Z days: the label YYYY-MM-DD 12:00 (UTC) covers 12:00 UTC of that day to 12:00 UTC of the next (21:00 to 21:00 Japan Standard Time) –

catchment_outlets.csv

One row per catchment, in the order of the spatial axis of the simulation.

Column Content
node_id identifier of the catchment
lon, lat location of the outlet, the centre of its pixel in the 1-arcsecond J-FlwDir grid (WGS84, degrees)
drainage_area_km2 area drained at the outlet, upstream catchments included
catchment_area_km2 area of the catchment itself
downstream_node_id the catchment it drains into; empty at the outlet of a river basin
is_basin_outlet the catchment is the outlet of one of the 710 river basins
has_gauge, has_dam a gauging station, a dam, is located at the outlet
evaluation_gauge one of the 979 gauges with quality-controlled discharge observations in D3P-JP
training_gauge one of the 332 gauges whose observations were used to train the model

The catchment polygons and the channel lines are in the network/ tables of D3P-JP.

Terms of use and citation

D3P-JP-Q is distributed under the CC BY-NC 4.0 licence; see LICENSE.md. Please cite the paper above, this record and the D3P-JP dataset.

Downloads last month
30