dg_index skipping in check_refine.m

[bhayward]

Hello all,

I was working on checking volume conservation in my Nested Grid setup using check_refine.m, and ran into an issue where there was a banding pattern appearing in several variables.

inverseDX.png (39.92 KiB) Viewed 289 times

inverseDY.png (36.54 KiB) Viewed 289 times

area.png (37.6 KiB) Viewed 289 times

bathy.png (122.24 KiB) Viewed 289 times

Looking at check_refine.m, the issue was coming from dg_index . I've only plotted the gradient in the eta direction, but there is a similar pattern in the xi direction.

gradIndex.png (17.58 KiB) Viewed 289 times

This comes from the way the Idg and Jdg points are set. It seems like Idg and Jdg should increase linearly, so I make them increase linearly.

Code: Select all

oldIdg = fix(XrF(Imid) - 0.5) + 1;
oldJdg = fix(YrF(Jmid) - 0.5) + 1;
Idg = (1:size(oldIdg,2))+oldIdg(1)-1;
Jdg = (1:size(oldJdg,2))+oldJdg(1)-1;

jdgOldNew.png (17.21 KiB) Viewed 289 times

This workaround seems to be helpful for me. This issue might be related to the lat/lon rho points of my grid, but I'm fairly certain I am correctly following the fine2coarse methodology.

inverseDX_new.png (33.72 KiB) Viewed 289 times

inverseDY_new.png (33.71 KiB) Viewed 289 times

area_new.png (61.15 KiB) Viewed 289 times

bathy_new.png (114.58 KiB) Viewed 289 times

I do apologize for the oversized images. I see there is an imgh and imgw option, but I can't seem to figure out where to put them in the text.

Thank you!

-Ben