MERIT
(Multi-Error-Removed Improved Terrain DEM) is a substantially improved
near-global terrain description with 90 m (3 arc-seconds) spatial resolution
(Yamazaki et al. 2017). MERIT covers almost all of Earth’s land areas within
90N-60S, except of Antarctica. Different to most other global DEM data sets,
MERIT provides – in good approximation – elevations of the bare ground. This
has been achieved by reducing vegetation heights (known as tree height bias)
using tree density and tree height maps as auxiliary information in the
production of the MERIT DEM. However, over built areas, MERIT elevations
may contain a bias due to urban canopy.MERIT relies on SRTM v2.1 South of 60°
latitude, ALOS AW3D North of 60° latitude, and uses elevations from Viewfinder
Panoramas (VFP-DEM) to fill voids (unobserved areas) where present. For the
void-filling with VFP-DEM, an average matching method has been applied by
Yamazaki et al. (2017) to ensure smooth transitions. Within the SRTM data area,
about 0.5 % of MERIT land cells rely on VFP-DEM. North of 60°, the contribution
of VFP-DEM is about ~30 %. As a result, elevation errors previously present in the SRTM model have
been reduced such that actual terrain features stand out more clearly.
Geodetic information: The MERIT DEMs are vertically referenced to the
EGM96 geoid and horizontally referenced to the WGS84 (World Geodetic System
1984).
Further notes: The MERIT DEM mostly represents bare ground
elevations, so is technically close to a digital terrain model (DTM). This makes the data set suitable for applications
requiring heights of the bare ground. Example areas are hydrology,
hydrodynamics, physical geodesy and geophysics.
Data access: The authors freely share their model for
non-commercial applications (e.g. science and education) via URL: hydro.iis.u-tokyo.ac.jp/~yamadai/MERIT_DEM/
References:
Yamazaki, D.,
D. Ikeshima, R. Tawatari, T. Yamaguchi, F. O’Loughlin, J.C. Neal, C.C.
Sampson, S. Kanae, P.D. Bates (2017), A high accuracy map of global terrain
elevations, Geophysical Research Letters, Doi: 10.1002/2017GL072874