David Sandwell
June 19, 2012


I discovered a bug in the code that merges the terrestrial gravity data from Canada and Alaska with the satellite-derived gravity.

Here is the bad code.

#
# update the arctic grid with Canadian Data
#
grdtrack can.xyz -GV19_ArcGP.grd -V > can.xyzs
grdtrack alaska.xyz -GV19_ArcGP.grd -V > alaska.xyzs
#
# make difference data
#
awk '{print $1, $2, ($3-4)}' < can.xyzs > can.xyd
awk '{print $1, $2, ($3-4)}' < alaska.xyzs > alaska.xyd
#

These two awk commands were suppose to remove column 4 which contains the satellite gravity but the $ is missing so the number 4 was subtracted instead.

These difference data were gridded and added back to the satellite gravity.  However, since the satellite gravity was never subtracted, the gravity values in these locations were basically doubled.  This effected all of Alaska and Canada and significant offshore areas.  

The anomalies looked OK but the values were 2X too large.

This has been corrected in V20.