id summary reporter owner description type status priority milestone component version severity resolution keywords cc branch_review mp_status wp_comment
613 Problem with ORCA2 coordinates at NEMO vn2+ idcatukmo nemo "Dear all,
The central ORCA2 download page at
http://www.nemo-ocean.eu/Using-NEMO/Configurations/ORCA2_LIM
says
""CAUTION 2 versions are availbable:
* ORCA2_LIM.tar could be used with all NEMO versions previous to
the tag nemo_v2 i.e. downloaded before 12th november 2006
* ORCA2_LIM_nemo_v2.tar is dedicated to NEMO versions tagged
nemo_v2 and the following ones nemo_v2_xx""
etc.
Can someone tell me the reason for this change please? I ask because the u,v and f coordinates are now wrong in row 148 (=last physical row of ORCA2). The T-points are OK, I think.
Attached is a polar plot of some ORCA2 points near the north pole.
The frame on the left is from the old (pre-vn2) ORCA2 coordinates file;
the one on the right is from the new (post-vn2) ORCA2 coordinates file.
Consider point (135, 148). In this region of the grid i is increasing towards the top of the page and j is increasing towards the left of the page. This can be shown by printing out the coordinates of T(134:136, 148), and T(135, 147:149).
(Note that, because ORCA2 uses a t-pivot north fold, row 148 ""doubles back on itself"", so that, for instance, T-point (49, 148) coincides with T-point (135, 148). The picture shows this by also highlighting point (52, 148), where the i and j indices are increasing in the downward and rightward directions, respectively.)
Anyway, for the old coords file (left frame), U(135, 148) and V(135,
148) are in the right locations wrt T(135, 148), and the four f-points F(134:135, 147:148) surround T(135, 148), as required. Similarly for point (52, 148).
On the new grid (right frame), however, U(135, 148) and V(135, 148) are wrongly located wrt T(135, 148) (they match the orientations of the grid around (52, 148), where the indices are in the reverse direction, on account of the the doubling back of the grid.) The f-points are even worse: they lie on a line, instead of surrounding T(135,148). (Note that F(134, 147) = F(135, 148); in fact we should have F(134, 147) = F(49, 148).)
So: can anyone say why the ORCA2 coords file was changed at NEMO vn2? (I notice that the scale factors e1{u,v,f} and e2{u,v,f} are different too.) Is it safe to keep using the old one? It matters to us because we wish to use conservative interpolation, and to calculate the remapping coefficients, we need the corners of the cells, ie the f-points.
Thanks for any help,
Ian.
" Bug closed low Unscheduled OCE release-2 fixed ORCA2 nemo_st@…