5.A.vi. DEM for Greenland

Primary Developers:		H. Jay Zwally and Robert A. Bindschadler (NASA Goddard Space Flight Center), Anita Brenner and John DiMarzio (Hughes STX) for the Oceans and Ice Branch of the Laboratory for Hydrospheric Physics, NASA/Goddard Space Flight Center.
Title:		DEM for Greenland from GEOSAT Altimetry
Publication Date:		1997
Bibliographic Citations:		National Snow and Ice Data Center (NSIDC), 1997. SEASAT and GEOSAT Altimetry Data for the Antarctic and Greenland Ice Sheets. University of Colorado, National Snow and Ice Data Center, Boulder, Colorado. (CD-ROM)
	**	Zwally, H. Jay, Anita C. Brenner, John DiMarzio, and Robert A. Birndschadler, 1997. DEM of Greenland from GEOSAT Altimetry (in NSIDC, 1997).
		Zwally, H. Jay, A.C. Brenner, J.A. Major, T.V. Martin, and R.A. Bindschadler, 1990. Satellite radar altimetry over ice. Volume 1, Processing and Corrections of SEASAT Data over Greenland. NASA Reference Publication 1:1233.
		Zwally, H. Jay, Judith A. Major, Anita C. Brenner, Robert A. Bindschadler, and Thomas V. Martin, 1990. Satellite radar altimetry over ice. Volume 2, Users' Guide for Greenland Elevation Data from SEASAT. NASA Reference Publication 2:1233.
Post-processing:		Jet Propulsion Laboratory (JPL) for the Multiangle Imaging Spectro Radiometer DEM.
Post-processing (Step 2):		National Oceanic and Atmospheric Administration (for GLOBE).
Bibliographic Citation for Post-processed DEM:	*	Bryant, Nevin, Richard Fretz, Niles Ritter, and Rafael Alanis, 1995. JPL/MISR 30 Arc-second Digital Elevation Model. Jet Propulsion Laboratory, Pasadena, California.
Source/Lineage Category:		12 (and partly 13)
Note:		The best current DEM for Greenland is a hybrid of this DEM, DTED, and DCW-based contours as described below and in Section 5.B.i.d.

There is only limited DTED coverage for Greenland. Where available, these data were used.

For other areas, several DEMs have been created using SEASAT, GEOSAT, and ERS-1 satellite altimetry. These data provide closely spaced spot measurements of the distance between the satellite platform carrying the altimeter. These measurements undergo various corrections and calibrations to produce spot measurements of elevation. Statistical processes may be used to reduce noise where measurement densities are sufficiently high. The resultant spot measurements are then interpolated into elevation grids (e.g. DEMs). The only version of such data that we have evaluated enough to use is this contribution based on GEOSAT data (see two paragraphs below). Further improvements of altimetry-based data are anticipated.

Alternatively, Operational Navigation Charts, upon which DCW was based, provide topographic contours south of about 72^oN latitude. North of that latitude, contours exist in some coastal areas, with inland areas having only a sparse collection of point elevations apparently collected on traverses. USGS used DCW to create its DEM of Greenland, except where DTED was available.

Another DEM was created from GEOSAT altimetry by Zwally and others (NSIDC, 1997; see list under "Primary Developers," above). This was provided to the MISR/JPL DEM project, in support of the NASA EOS Multiangle Imaging SpectroRadiometer mission, by GLOBE Task Team member J.-P. Muller. The original DEM had 10 km gridding, but was resampled to 30" by JPL for use in the MISR mission. (Further documentation is available from NSIDC at http://www-nsidc.colorado.edu.) North of 72^oN latitude, JPL used Digital Chart of the World contours to create a gridded DEM.

Cell-centered registration at lower resolution; value replicated from 10km grid for Greenland. Graphic describing georeferencing and sampling for the Zwally (and others)/NSIDC/JPL data for Greenland.

Plate 20. Plate 21. Click on each image to view larger size.

Plate 20 only incorporates data for the area of the original Zwally (and others)/NSIDC/JPL DEM ultimately used in GLOBE. As that area is dominated by the Greenland Ice Sheet, it is not surprising that the histogram is dominated by elevations of 2000 - 3000m. Note that this DEM is dissimilar from other DEMs in GLOBE, as are its sources. Further assessment may follow in future versions of GLOBE, after other candidate DEMs from altimetric sources are assembled and analyzed. Plate 21 shows the two-degree wide blend zone between the data contributed by JPL and those from DCW and DTED in Greenland. The histogram shows that dominant elevations are about 2000m in this area off the center of the ice cap.

NGDC's assessment of NIMA DTED, USGS, and Zwally (and others)/NSIDC/JPL DEMs suggested that NIMA DTED were preferred where available. The USGS model was preferred in many coastal areas in the south, southeast and southwest, where DCW hypsography was relatively detailed. However, in the interior, even south of 72^o latitude where DCW contours were fairly detailed, the Zwally (and others)/NSIDC/JPL DEM appeared preferable.

It was suggested by J.-P. Muller of University College London (Geomatic Engineering Department) that even where DCW contours existed on the Greenland Ice Sheet, that they were not necessarily very accurate compared to the NSIDC model. Inspection at NGDC (now NCEI) for terrain-like characteristics in the two DEMs supported this assessment. The Zwally (and others)/NSIDC/JPL DEM was thus preferred primarily at higher latitudes and inland. At these latitudes, this DEM appeared to overrun coastlines in some areas, perhaps because the satellite altimetry gave readings on ice features (and thus considered land by that model, whereas World Vector Shoreline (WVS) coastlines considered some of these areas to be oceanic).

Therefore, NGDC (now NCEI) mosaicked a DEM for Greenland based on the following priorities:

1. DTED-based 30" data where available.
2. USGS/GTOPO30 data based on DCW, where DCW contained sufficient detail (generally southerly near-coastal areas not characterized by substantially permanent ice).
3. Zwally (and others)/NSIDC/JPL data in other areas (generally in the interior of Greenland, and in northerly coastal areas characterized by substantially permanent ice).
4. NGDC (now NCEI) performed a linear transition between the Zwally (and others)/NSIDC/JPL DEM and the DCW-based DEM, over a two-degree wide zone. This transition is category 13 in the source/lineage map.
5. In some northern coastal areas characterized by substantially permanent ice, USGS/GTOPO30 values sometimes extended oceanward of the limit of Zwally (and others)/NSIDC/JPL coverage. In this case, the USGS/GTOPO30 data were used. In some other areas, Zwally (and others)/NSIDC/JPL values sometimes extended oceanward of USGS/GTOPO30 data coverage, which was limited by a selection of DCW and WVS coastlines. In such cases, the coastlines used by USGS/GTOPO30 were accepted, thus masking out Zwally (and others)/NSIDC/JPL elevations in areas considered oceanic by USGS/GTOPO30.