Open remkos opened 9 years ago
For altimetry, we will need a model of geoid change (dGeoid). Peltier doesn't appear to provide an ICE6G dGeoid dataset. The available datasets (dsea) are appropriate for tide gauges, but include vertical land motion that we don't want for altimetry. Because dGeoid is significant only at long wavelengths, it's likely that any dGeoid changes between ICE6G and ICE5G are modest.
The dGeoid grid can easily be computed by adding dsea and drad. Both grids are provided for ICE6G.
I have determined the dGeoid for ICE6G, see the image below.
The difference between dGeoid for ICE6G and ICE5G is still significant, as shown in the image below. Differences go up to ±0.1 mm/year over ocean, while the over-ocean maximum of dGeoid is about 0.5 mm/year.
For comparisons with GRACE, I'd prefer to have dgeoid from the A, Wahr, and Zhong GIA model. The dgeoid field provided on Tellus (ftp://podaac-ftp.jpl.nasa.gov:/GeodeticsGravity/tellus/L3/pgr/GIA.COMPRESSIBLE.GRIDS.nc) isn't correct for altimetry. With this advice from Mark Tamisiea, I've created a proper dgeoid field:
"Solve for the 2,1 components of the geoid result and multiply by 2.06 (eq. A7 in Tamisiea 2011). Then find the sum over the global oceans for the geoid model (with the increased 2,1 values), G. Find the sum over the global oceans for the uplift model, R. Find constant M such that G + M = R. Add M back to the geoid model (a new degree 0 value)."
It has significant differences with the Peltier ICE6G field:
Perhaps we should include both options.
Peltier produced a new GIA model, called ICE6G. Apparently this appears to be a significant improvement for present-day GIA ... http://www.atmosp.physics.utoronto.ca/~peltier/data.php