Custom Query (2547 matches)
Results (31 - 33 of 2547)
Ticket | Resolution | Summary | Owner | Reporter |
---|---|---|---|---|
#1059 | fixed | Wrong scale factors with key_vvl in dynspg_flt.F90 and bdydyn.F90 | jchanut | jchanut |
Description |
With key_vvl only:
spgu(ji,jj) = spgu(ji,jj) + fse3u(ji,jj,jk) * ua(ji,jj,jk) spgv(ji,jj) = spgv(ji,jj) + fse3v(ji,jj,jk) * va(ji,jj,jk) it should rather be: spgu(ji,jj) = spgu(ji,jj) + fse3u_a(ji,jj,jk) * ua(ji,jj,jk) spgv(ji,jj) = spgv(ji,jj) + fse3v_a(ji,jj,jk) * va(ji,jj,jk)
Barotropic velocity is computed (then removed) from ua/va arrays again with "now" vertical scale factors: pu2d(:,:) = pu2d(:,:) + fse3u(:,:,jk) * umask(:,:,jk) * ua(:,:,jk) pv2d(:,:) = pv2d(:,:) + fse3v(:,:,jk) * vmask(:,:,jk) * va(:,:,jk) ... pu2d(:,:) = pu2d(:,:) * phur(:,:) pv2d(:,:) = pv2d(:,:) * phvr(:,:) again fse3u_a/fse3v_a arrays should be used, as well as inverse depths at next time step. |
|||
#365 | fixed | wrong scale factors used in bdyvol | nemo | gm |
Description |
in bdyvol.F90 the sum of the forcing contribution is performed using e1v * e2u whereas the surface of a box at T-point is simply e1t*e2t. Therefore the folling lines:
should be changed into
or in a more compact form (like the one used in obcvol.F90)
by the way, introducing a scalar equal to 1/rauw will reduce the cpu time (less division):
gurvan |
|||
#507 | fixed | wrong scale factors in VLL | nemo | rblod |
Description |
Fix:
Thanks to Dave and Rachel for addressing this. |