The Global Land One-km Base Elevation (GLOBE) Project
Candidate Review: DEM for Japan
The GLOBE project has three digital elevation models (DEMs) for Japan to choose from for the final 1-km global digital elevation model. The three candidate DEMs were contributed by the Geographical Survey Institute of Japan (GSI), the U.S. Geological Survey EROS Data Center (EDC), and the National Imagery and Mapping Agency (NIMA). The results of the evaluation of the three DEMs are found below.
The first section includes a table showing when each DEM equaled the minimum, maximum, and median of the three DEMs.
The second section displays histograms created from each digital elevation model.
We are asking for input from the GLOBE partners and anyone else who is interested in analyzing the results we have obtained from the three models so far. Although review for GLOBE Version 1.0 is essentially complete, review for future versions of GLOBE remains open. Please telephone or email Paula Dunbar with any comments.
MINIMUM, MAXIMUM, and MEDIAN HITS
A table listing the results of an evaluation of when each DEM equaled the minimum, maximum, and median of all three DEMs is listed below. It can be seen from this table that the median values (244595 hits) and the maximum values (300459 hits) occur most often found in the EDC DEM and the minimum values (428609 hits) occur most often in the NIMA DEM.
In addition, the same values occur most often in the NIMA and EDC DEMs (58016 hits) and least often in the NIMA and GSI DEMs (13800 hits). The similarity of the NIMA and EDC DEMs is not surprising, since these DEMs were derived from the same source materials (3-arc-second DTED data).
To view histograms from the three DEMs for Japan click on the icons listed below.
Spikes at 100 m, 200 m, 300 m ... 1000 m, are evident in the histograms from the EDC and NIMA DEMs. This terracing effect, where certain height values occur in a model much more frequently than other height values, often results when terrain models are derived from topographic maps. The large peaks often correspond to contour intervals on the source maps. The source data for both of these DEMs was the former Defense Mapping Agency (DMA) Digital Terrain Elevation Data (DTED). DTED are developed on a 3-arc-second latitude longitude grid, showing elevations above sea level based on the World Geodetic System 84 datum. Source materials for DTED may include digital elevation models, sterographic imagery, or topographic maps. The terracing effect displayed in both histograms indicates the source materials for these DTED data were probably topographic maps. The differences in the two DEMs are due to differences in how the DTED data were resampled to produce the final 30-arc-second DEMs.
Contrasted with the EDC and NIMA histograms, the histogram for the GSI DEM shows a very smooth curve with no peaks. The source for this DEM was an adaptation of GSI's 250 m gridded DEM. The GSI 250 m DEM was in a projection and datum used in Japan. It was reprojected to the World Geodetic System 84 datum and a 30-arc-second latitude longitude grid, as a contribution to the GLOBE project.
COLOR PALETTE IMAGES
To view color palette images for Hokkaido and the Tokyo, Honshu area, from the three Japanese DEMs click on the icons listed below.
Color palette images for Hokkaido:
Color palette images for the Tokyo, Honshu area:
SHADED RELIEF IMAGES
To view shaded relief images for Hokkaido and the Tokyo, Honshu area, from the three Japanese DEMs click on the icons listed below.
Shaded relief images for Hokkaido:
Shaded relief images for the Tokyo, Honshu area:
The level of detail as displayed in the shaded relief images is very similar in all three DEMs.
To view aspect maps for Hokkaido and the Tokyo, Honshu area, from the three Japanese DEMs click on the icons listed below.
Aspect maps for Hokkaido:
Aspect maps for the Tokyo, Honshu area:
The level of detail as displayed in the aspect maps is very similar in all three DEMs. Although artifacts can be found in the three images when rigorously compared, the quality of the three DEMs appears to be very similar.