MERIT Hydro · 90 m · Downloads

README.txt

MERIT Hydro · 90 m
==================

Products computed on the MERIT Hydro flow directions.

Grid        one global grid, 432,000 x 174,000 pixels at 3 arc-seconds,
            WGS 84 (EPSG:4326)
Source      MERIT Hydro (Yamazaki et al., 2019, https://doi.org/10.1029/2019WR024873)
            on MERIT DEM (Yamazaki et al., 2017, https://doi.org/10.1002/2017GL072874)
            Register on https://global-hydrodynamics.github.io/MERIT_Hydro/
            (CC BY-NC 4.0 / ODbL) and cite Yamazaki et al. (2019).
            The MERIT Hydro layers themselves (dir, elv, upa, upg, wth, hnd)
            are not redistributed here.

Tools with products here
  flowdivide/   v1.0, v1.1
  flowtopo/     v1.0 (on Zenodo), v1.1 (in preparation)
  flowmorph/    v1.0 (on Zenodo), v1.1 (in preparation)

Dataset paper  MERIT-FullBasin, ESSD discussion, https://doi.org/10.5194/essd-2026-402
Archive        version 1.0 on Zenodo, https://doi.org/10.5281/zenodo.20344113
               100 files, 142.2 GB, one zip per region, stored uncompressed so
               GDAL can read one file inside a zip without downloading it:

    export GDAL_HTTP_USERAGENT="your-tool/1.0"   # Zenodo answers 403 without one
    ZIP=/vsicurl/https://zenodo.org/api/records/20344113/files/region_00064.zip/content
    TIF=dfs_interval/dfs_in/dfs_in_region_00064_90m.tif
    gdallocationinfo -valonly "/vsizip/{$ZIP}/$TIF" 16200 13200

Every file of the dataset, with its size, is listed in Table 1 on
https://fullhydro.org/merit-90m/v1.1/fullbasin/#files

Name	Size	Description
../		up one level
flowdivide/	1.6 GB	FlowDivide on MERIT Hydro · 23 files
flowtopo/		FlowTopo on MERIT Hydro · no files yet
flowmorph/		FlowMorph on MERIT Hydro · no files yet
README.txt	1.5 kB	what this folder holds

The same folder on the data server: https://data.fullhydro.org/merit-hydro-90m/

Table 1 of the paper

Every file in the dataset

This is Table 1 of the paper, with two columns added: where each file can be downloaded. All native-resolution rasters share one global grid, 432,000 × 174,000 pixels at 3 arcsec. The file-name wildcard * is the region_id, written with five digits (region_00064).

Read the compressed column, not the raw one. Nothing is distributed at its uncompressed size — that column is there to show what the compression is doing.

File	Variable	Type	NoData	Compressed	Raw	Here	Zenodo
Input baseline topology — MERIT Hydro, not redistributed
dir_global_90m.tif	Flow direction	UInt8	247	~6.5 GB	~70 GB	✗	✗
upg_global_90m.tif	Flow accumulation in pixels	UInt32	0	~18 GB	~280 GB	✗	✗
upa_global_90m.tif	Flow accumulation in km²	Float32	−9999	~36 GB	~280 GB	✗	✗
Basin and region masks with attribute tables — global
bsn_global_90m.tif	Basin mask (the archive's DEFLATE GeoTIFF; the COG on this site is 251 MB)	UInt32	0	~0.8 GB	~280 GB	✓	✓
rgn_global_90m.tif	Region mask	UInt8	0	~0.1 GB	~70 GB	✓	✓
bat_global_90m.csv	Attribute table, one row for each of the 24,587,290 outlets. Here: only the 340,991 basins of at least 1 km².	CSV	—	4.3 GB	—	✓	✓
rat_global_90m.csv	Region attribute table	CSV	—	10 kB	—	✓	✓
Basin and region masks with attribute tables — per-region tiles
bat_region*_90m.csv	Basin attribute table	CSV	—	6.3 GB	—	✗	✓
bsn_region*_90m.tif	Basin mask	UInt32	0	~0.85 GB	~280 GB	✗	✓
bsn_region*_90m_remapped.tif	Basin mask renumbered with the local basin id	UInt32	0	~0.85 GB	~280 GB	✗	✓
DFS interval index — per-region tiles
dfs_in_region*_90m.tif	DFS entry timestamp	Int32	−1	~40 GB	~280 GB	✗	✓
dfs_out_region*_90m.tif	DFS exit timestamp	Int32	−1	~40 GB	~280 GB	✗	✓
Upstream morphometric attributes — per-region tiles, populated where upstream area reaches 10 km²
mean_elev_upstream_region*_90m.tif	Mean upstream elevation	Float32	−9999	~3 GB	~280 GB	✗	✓
min_elev_upstream_region*_90m.tif	Minimum upstream elevation	Float32	−9999	~3 GB	~280 GB	✗	✓
max_elev_upstream_region*_90m.tif	Maximum upstream elevation	Float32	−9999	~3 GB	~280 GB	✗	✓
std_elev_upstream_region*_90m.tif	Standard deviation of upstream elevation	Float32	−9999	~3 GB	~280 GB	✗	✓
mean_slope_upstream_region*_90m.tif	Mean upstream slope	Float32	−9999	~3 GB	~280 GB	✗	✓
std_slope_upstream_region*_90m.tif	Standard deviation of upstream slope	Float32	−9999	~3 GB	~280 GB	✗	✓
perimeter_catchment_region*_90m.tif	Catchment perimeter	Float32	−9999	~3 GB	~280 GB	✗	✓
length_basin_region*_90m.tif	Basin length	Float32	−9999	~3 GB	~280 GB	✗	✓
width_basin_region*_90m.tif	Basin width	Float32	−9999	~3 GB	~280 GB	✗	✓
convexity_planform_region*_90m.tif	Planform convexity	Float32	−9999	~3 GB	~280 GB	✗	✓
relief_basin_region*_90m.tif	Basin relief	Float32	−9999	~3 GB	~280 GB	✗	✓
hypsometric_integral_region*_90m.tif	Hypsometric integral	Float32	−9999	~3 GB	~280 GB	✗	✓
elongation_ratio_region*_90m.tif	Elongation ratio	Float32	−9999	~3 GB	~280 GB	✗	✓
compactness_gravelius_region*_90m.tif	Gravelius compactness	Float32	−9999	~3 GB	~280 GB	✗	✓
circularity_ratio_region*_90m.tif	Circularity ratio	Float32	−9999	~3 GB	~280 GB	✗	✓
lemniscate_ratio_region*_90m.tif	Lemniscate ratio	Float32	−9999	~3 GB	~280 GB	✗	✓

Added on this site

Five files that are not in the paper. They are cuts and vector forms of what is already in Table 1, made so the data can be opened without a raster workflow. The outlines are simplified — at full resolution they would come to 6.9 GB. They will go into the Zenodo archive with the next release.

File	What it is	Type	Size	Here	Zenodo
bat_global_90m_gt1km2.csv	The basin table cut to the 340,991 basins of at least 1 km², all 27 columns	CSV	60 MB	✓	v2
bsn_global_gt1km2.gpkg	340,955 basin polygons, simplified, holes kept	GeoPackage	143 MB	✓	v2
bnd_global_gt1km2.gpkg	Outer boundary of those basins, as lines, no holes	GeoPackage	141 MB	✓	v2
rgn_global.parquet	The 90 region outlines, simplified	GeoParquet	5 MB	✓	v2
rat_global_90m_with_tiles.csv	The region table plus the MERIT Hydro tile list for each region	CSV	31 kB	✓	v2

Level one · 90 rows

Regions

Every region box is a whole number of degrees. At 3 arcsec one degree is exactly 1,200 pixels, so a region’s row and column offsets are always multiples of 1,200. That is why a region can be cut straight out of the global rasters with integer arithmetic — gdal_translate -srcwin and nothing else — with no resampling and no half-pixel shift.

IDs: two digits for a region that is a whole HydroBASINS Level-02 unit (57 regions, 11–92), four for a part of one (3513, 3546, 5618, 5699, 7711, 7724, 9112, 9134: the Level-02 unit, then the last digits of the first and last Level-03 unit in that part, so 9134 holds 913 and 914), 10001–10002 for the two groups cut by the antimeridian, and 20000 + continent × 100 + n for the 23 island groups (20101–20901). For a continental region the first digit is the continent; for an island group it is the digit after 20.

How the 90 regions were formed

Regions follow the manually drawn HydroBASINS units. Each basin is assigned to a Level-03 unit by area overlap; the Level-03 children are then merged back into their Level-02 parent, but only where the merged region still fits the capacity of 1400 square degrees. Parents that fail that test are split into parts, each a run of consecutive Level-03 units. Basins that cross the antimeridian and isolated islands are grouped on their own. The result is 90 regions: 65 continental — 57 whole Level-02 units (IDs 11–92) and 8 parts of the four units that had to be split (3513, 3546, 5618, 5699, 7711, 7724, 9112, 9134) — two antimeridian groups (10001–10002) and 23 island groups (20101–20901, the first digit after 20 being the continent).

Level two · 24,587,290 rows

Basins

The raster and the table share one address. A basin ID read out of the mask is table_index + 1: basin 1 is row 0, basin 2 is row 1, and so on. Going from a pixel to its record is an array offset, never a join and never a lookup table. Rows are sorted by area, largest first, which is why basin 1 is the Amazon.

The full table has one row for every outlet, 13.3 million of them single coastal pixels. The basin table in flowdivide/ carries the 340,991 basins of at least 1 km², which together cover more than 99.5% of the land.

The outlines are simplified; the numbers are not. Every attribute column — area, outlet, bounding box — is computed on the native 3 arcsec grid and is not measured from these outlines. When the boundary itself matters, take the 90 m rasters.

Original source data

Based on MERIT DEM (Yamazaki et al., 2017) and MERIT Hydro (Yamazaki et al., 2019).

Register on the MERIT Hydro page (CC BY-NC 4.0 / ODbL) and cite Yamazaki et al. (2019). We do not redistribute its baseline layers — dir, elv, upa, upg, wth, hnd — so those come from the MERIT Hydro team.

Reuse

How to cite

Dataset Jiang, L. et al. (2026). MERIT-FullBasin: a global 90 m basin dataset with a per-pixel upstream index and morphometric attributes. Zenodo. 10.5281/zenodo.20344113. Upstream index under ODbL 1.0, other layers CC BY 4.0; published 22 May 2026.

Paper Jiang, L. et al. (2026). Earth System Science Data, under review. 10.5194/essd-2026-402

The MERIT Hydro layers this is built on are not redistributed here. Cite them separately: Yamazaki et al. (2019), MERIT Hydro: a high-resolution global hydrography map based on latest topography datasets, Water Resources Research.