I would like to compare the accuracy of the SRTM 90m resolution DEM and the ASTER 30m resolution DEM for Belize, by comparing the elevation values against spot heights taken from maps of the country. I have georeferenced tiff files of 1:50,000 map sheets covering Belize, which include spot heights. I projected the files to the WGS84 lat/long coordinate system to match the DEM files I downloaded. The map legend states that elevation is measured in metres, using Clarke 1866 Spheroid, Vertical Datum Mean Sea Level. I have created a Point layer and randomly picked 50 spot heights spread across the country from the map sheets. I then used the 'Extract Values to Points' tool to obtain the heights from the SRTM and ASTER layers for the 50 spot height coordinates I'd picked. However, I understand from the metadata for SRTM and ASTER that the WGS84 elevations are likely to be measured from the EGM96 geoid. Before I compare the elevation values (which I plan to do in SPSS using Paired t-test analyses), I think I need to convert them so they're all measured from the same geoid. Does anyone know how I may be able to do this? Any help will be greatly appreciated! Thanks in advance. Fliss
You're dealing with two datum differences - horizontal datum and vertical.
Make sure you use the appropriate transformation of your test points to project your horizontal locations properly to the WGS84 datum. You didn't say what your horizontal datum was on your maps; NAD27 is based on Clarke_1866 spheroid so maybe that will do. If so, I believe the appropriate transformation would be "NAD_1927_To_WGS_1984_2" (ArcGIS install folder/Desktop10.0/Documentation/geographic_tranformations.pdf, p. 22).
Vertical transformations are another deal - you'll need to find out what local sea level is in EGM96 and offset your sea-level datum Z values accordingly.