Ming's initial research identified the SRTM Level 4 DEM data distributed by the CGIAR Consortium (http://srtm.csi.cgiar.org)
as best suited to this project's requirements.

During our teleconference on 4/12 Brian McGill described some of the Terrain Analysis that will use the fused DEM layer
as a primary input. These analyses will be very sensitive to the terrain 'signal' in the DEM layer.

The CGIAR Level 4 DEM is a 'refined' terrain model product; the original SRTM DEM is subjected to a multi-step enhancement
process.

Refer to : http://srtm.csi.cgiar.org/SRTMdataProcessingMethodology.asp for details

This refinement processing raises two issues of concern in my mind:

1) There is a possibility that these processing steps have reduced the quality of the terrain 'signal' that is most
relevant to some or all of the planned terrain analysis. The magnitude of the reduction may vary for each
of the terrain analysis tasks.

2) The processing may cause a significant discontinuity along the edges between the CGIAR/SRTM tiles and
the ASTER GDEM tiles used to represent high-latitude regions.

Prior to final acceptance of the CGIAR Level 4 DEM as one of the bases (ASTER being the other) for the fused terrain
data layer, I recommend that the project scientists review the CGIAR/SRTM data processing methodology within the
context of planned terrain analysis, and that the constructors of the fused terrain layer evaluate the 'boundary
edge' regions for a sampling of locations.

Following this review, all concerned can re-confirm that the CGIAR/SRTM product is the best choice for this project,
and we can proceed with full-scale production of the fused global terrain layer.