MERIT Hydro: global hydrography datasets

Last Update: 17 May, 2019


General Information

Product Info

MERIT Hydro is a global hydrography datasets, developed based on the MERIT DEM and multiple inland water maps. It contains flow direction, flow accumulation, hydrologically adjusted elevations, and river channel width.
The dataset can be downloaded after the registration & license agreement by Google Form.(see below)

Data Summary

High-resolution raster hydrography maps are a fundamental data source for many geoscience applications. Here we introduce MERIT Hydro, a new global flow direction map at 3 arc-second resolution (~90 m at the equator) derived from the latest elevation data (MERIT DEM) and water body datasets (G1WBM, GSWO, and OpenStreetMap). We developed a new algorithm to extract river networks near-automatically by separating actual inland basins from dummy depressions caused by the errors in input elevation data. After a minimum amount of hand-editing, the constructed hydrography map shows good agreement with existing quality-controlled river network datasets in terms of flow accumulation area and river basin shape. The location of river streamlines was realistically aligned with existing satellite-based global river channel data. Relative error in the drainage area was smaller than 0.05 for 90% of GRDC gauges, confirming the accuracy of the delineated global river networks. Discrepancies in flow accumulation area were found mostly in arid river basins containing depressions that are occasionally connected at high water levels and thus resulting in uncertain watershed boundaries. MERIT Hydro improves on existing global hydrography datasets in terms of spatial coverage (between N90 and S60) and representation of small streams, mainly due to increased availability of high-quality baseline geospatial datasets. The new flow direction and flow accumulation maps, along with accompanying supplementary layers on hydrologically adjusted elevation and channel width, will advance geoscience studies related to river hydrology at both global and local scales.


MERIT Hydro river network for Tigris-Euphrates, Ob, Amazon, and Pearl Rivers.
High resolution image can be downloaded from "Quick Look Figure" below.

MERIT Hydro world tour on Google Earth. (c) George Allen (Texax AM Univ.)

Use of Data

How to download

Password for download is issued after the user resistration. Please carefully read the below agreement and register on Google Form.

License Agreement

The MERIT Hydro is licensed under a Creative Commons "CC-BY-NC 4.0" or Open Data Commons "Open Database License (ODbL 1.0)".
(i.e. dual license, you can choose an appropriate license for you)

To view a copy of these license, please visit:
CC-BY-NC 4.0 license: Non-Commercial Use with less restriction.
ODbL 1.0 license: Commertial Use is OK, but the derived data based on MERIT Hydro should be made publicly available under the same ODbL license.
For example, if you create a flood hazard map using MERIT Hydro and you'd like to provide a COMMERCIAL service based on that, you have to make the hazard map PUBLICLY AVAILABLE under OdBL license.

Note that the above license terms are applied to the "derived data" based on MERIT Hydro, while they are not applied to "produced work / artwork" created with MERIT Hydro (such as figures in a journal paper). The users may have a copyright of the artwork and may assign any license, if when the produced work is not considered as "derived data".

By downloading and using the data the user agrees to the terms and conditions of the license. Notwithstanding this free license, we ask users to refrain from redistributing the data in whole in its original format on other websites without the explicit written permission from the authors.

MERIT Hydro is available for download at
The copyright of MERIT Hydro is held by the developers, 2019, all rights reserved.

Data sharing policy

Citation to the paper is adequate if you simply use the MERIT Hydro. If you asked help for additional handling/editing of the dataset, or your research outcome highly depends on the product, the developer would request co-authorship.

[ Citation ]
Yamazaki D., D. Ikeshima, J. Sosa, P.D. Bates, G.H. Allen, T.M. Pavelsky
MERIT Hydro: A high-resolution global hydrography map based on latest topography datasets
Water Resources Research, ACCEPTED, 2019, doi: 10.1029/2019WR024873
-- Absstract in WRR

MERIT Hydro (available at

Registration for Download

Please fill the Google Form for registration & license agreement. The password for downloading is emailed after registration.

or please contact to the developer (yamadai [at] to get an access.

Data Description

Data Format

The flow direction datasets is referenced to WGS84 system and the EGM96 geoid.
The data is prepared as 5 degree x 5 degree tiles (6000 pixel x 6000 pixel).
Filename represents the center of the lower left pixel of the data domain; e.g. the file "n30w120_dir.tif" is for the Flow Direction of the domain N30-N35, W120-W115. (more accurately, N29.99958333-N34.99958333,W120.0004167-W115.0004167)

The 5 deg tiles are compressed into 30 degree (lat) by 30 degree (lon) package.
Package name represents the lower left corner of the package domain; e.g. the package "dir_n30w120.tar" contains the files in the domain of N30-N60, W120-W090.

The data format is geoTiff.


Flow Direction (Local Drainage Direction)

Flow direction is prepared in 1-byte SIGNED integer (int8). The flow direction is represented as follows.
1: east, 2: southeast, 4: south, 8: southwest, 16: west, 32: northwest, 64: north. 128: northeast
0: river mouth, -1: inland depression, -9: undefined (ocean)
NOTE: If a flow direction file is opened as UNSIGNED integer, undefined=247 and inland depression=255

Hydrologically Adjusted Elevations

Adjusted elevation is reprepared in 4-byte float (float32). The elevations are adjusted to satisfy the condition 'downstream is not higher than its upstream' while minimizing the required modifications from the original DEM.
The elevation above EGM96 geoid is represented in meter, and the vertical increment is set to 10cm. The undefined pixels (oceans) are represented by the value -9999.
For detailed method, see [Yamazaki et al., 2012, WRR].

Upstream Drainage Area (Flow Accumulation Area)

Drainage area is reprepared in 4-byte float (float32). The drainage area is represented in km2.
The undefined pixels (oceans) are represented by the value -9999.

Upstream Drainage Pixel (Flow Accumulation Grid)

Number of upstream pixel is reprepared in 4-byte integer (int32).
The undefined pixels (oceans) are represented by the value -9999.

River Channel Width

River width is reprepared in 4-byte float (float32), in meter. The values larger than 0 represents the river width at the channel centerlines. The value "-1" represents non-centerline water pixels, and the value "0" corresponds to the non-water pixels.
The undefined pixels (oceans) are represented by the value -9999.
River channel width is calculated by the method described in [Yamazaki et al. 2012, WRR], with some improvements/changes on the algorithm.

HAND: Height Above Nearest Drainage

HAND is reprepared in 4-byte float (float32), in meter. The values represent the height above nearest drainage [Noble et al., 2011, JoH]. Here, "drainage" is defined as pixels with upstream area >0.5km2.
The undefined pixels (oceans) are represented by the value -9999.

Quick Look Figure

About Figure Archive (password not required)

In order to quick look the product, the figures visualizing the river netowkrs are acceesible without password.
Each archive contains 5deg*5deg tile figures for the entire globe.
Please download the figure archive from the links below if you are interested in.

Download Figure Archive

River Network: flwdir.tar.gz
River Width: width.tar.gz
HAND: hand.tar.gz

NOTE: These figures are distributed under CC-BY 4.0 license.
Copyright of the figures held by Dai Yamazaki, 2019-.

Display on Google Earth

Figures can be displayed in Google Earth Pro (desktop app).
Caution: the Google Earth app might be crashed due to the KMZ file size is very big (~450MB).
Download: River Network KMZ
Download: River Width KMZ


- Supplement Data Layers (left: river width, right: HAND)


- Google Earth display image


Current version is v1.0.1 [10 June, 2019].

THe password is issued after registration on Google Form. (see above)


[v1.0.1] (10 Jun, 2019)Added some tiles which were missing due to GDAL bug in v1.0 [n50w085_upa.tif, n40e075_upg.tif, n60w130_upg.tif, n10w010_upg.tif].
[v1.0] (17 May, 2019)Official Release MERIT Hydro. River width around coastal area was modified, due to uncertainty in water masks in coastal zone.
[v0.7] (25 Jan, 2019)Pre-release version.

Flow Direction Map (password required)

Adjusted Elevation (password required)

Upstream Drainage Area (password required)

Upstream Drainage Pixel (password required)

River Width (password required)

HAND: Height Above Nearest Drainage (password required)

Known Problems

River Width

[Update from GWD-LR v1] Width algorithm was updated to consider sub-pixel water fraction. Now, 30m water map is used for width calculation at 90m resolution.
Currently river width is calculated for each channels separatelly in braided/anabranching sections. Merged river width should be calculated.

Water Body Map

There are some inconsistency between the DEM land sea mask and wate body data (such as new islands along the coast). The quality of OpenStreeetMap water body layer is not uniform iu all areas.

Channel Bifurcations

Channel bifurcation is not well represented in the current version. Each pixel is assumed to have only one downstream direction. Secondary (or multiple) downstream direction should be considered, to represent complex river networks in the delta regions, floodplains, and braided rivers.

Underground rivers/tunnels

Major underground rivers/tunnels should be implemented to improve large-scale water balance.

River/Lake Separation

Rivers and lakes are vetter to be separated for some applications.

Below sea level areas

The areas below sea level in coastal regions are not well represented in adjusted elevation data.

Flow direction over glacier

Flow direction over glacier is not well represented, because the elevation of glacier centerline is higher than glacier edge.

Supplementary Data

It's better to add: location of GRDC gauging stations, water falls, reservoirs, etc.

Data Sources

MERIT Hydro was developed by processing the following products as baseline data (all are freely available from their web page).

Digital Elevation Model

- U-Tokyo MERIT DEM: link to MERIT DEM webpage

Water Body Data

- U-Tokyo G1WBM water body data: link to G1WBM webpage
- OpenStreetMap water body layer: link to the OSM Water Layer webpage
- EC-JRC Global Surface Water Occurence: link to the GSWO

Other Input Data

- U-Maryland Landsat forest cover data: link to the Global Forest Change webpage