Support for Exchange between Faces #88
Conversation
|
This is currently very close. I just need to make it compatible with @jbusecke's grid generation module. |
|
Just updating @maikejulie and @jetesdal that my latest commit (85cb3e2) enables the ability to interpolate and difference 2D c-grid vectors (e.g. u and v) from cell edges to cell centers. This is what you need in order to do the volume and tracer budgets. What does not yet work is interpolating / differencing such vectors to the cell edges, such as for calculating vorticity. For various technical reasons, that turns out to be much more difficult. The testing is pretty bare bones. And I definitely need to add more documentation. I am going to write a bit more documentation and then merge this, hopefully within a few days. I want to get this into your hands so I can get your feedback about the API and help finding bugs. I just wanted to provide some evidence that I am, as promised, continuing to work on this (instead of sleeping). |
|
Should I wait for fixing the grid generation stuff, until the interpolation works? I am asking since that is an essential part of what my module uses. |
|
@jbusecke I don't understand your question. |
|
My bad, I did not read carefully. I will take a look at it in the next days and see if I can get the grid generation to work. |
|
Hey Ryan and Julius,
Sounds great, and I'm very glad we're discussing this! Having something
work for tracer and volume budgets is a great start.
I'll provide more feedback and help in the next few weeks. I'd be super
excited to have this work for vorticity budgets!
Thanks,
Maike
…On Wed, Jan 24, 2018 at 11:23 PM, Ryan Abernathey ***@***.***> wrote:
Just updating @maikejulie <https://github.com/maikejulie> and @jetesdal
<https://github.com/jetesdal> that my latest commit (85cb3e2
<85cb3e2>)
enables the ability to interpolate and difference 2D c-grid vectors (e.g. u
and v) from cell edges to cell centers. This is what you need in order to
do the volume and tracer budgets.
What does not yet work is interpolating / differencing such vectors to the
cell edges, such as for calculating vorticity. For various technical
reasons, that turns out to be much more difficult.
The testing is pretty bare bones. And I definitely need to add more
documentation. I am going to write a bit more documentation and then merge
this, hopefully within a few days. I want to get this into your hands so I
can get your feedback about the API and help finding bugs.
I just wanted to provide some evidence that I am, as promised, continuing
to work on this (instead of sleeping).
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#88 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AEY4jkw_vzL98xFwjLbxYtLDUAobwuxuks5tOAHXgaJpZM4QqzE->
.
|
|
The vorticity budgets will be super hard. This is due to the topology of the llc grid. Probably easier to discuss in person at ocean sciences. |
|
Definitely better in person. Scribbles are semi-essential.
Being able to compute the curl, streamfuctions etc are things that come up
regularly talking about the needs of the llc/ecco users, so I rate this as
important beyond just my own curiosity.
We should discuss.
Maike
…On Fri, Feb 2, 2018 at 2:25 PM, Ryan Abernathey ***@***.***> wrote:
The vorticity budgets will be super hard. This is due to the topology of
the llc grid. Probably easier to discuss in person at ocean sciences.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#88 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AEY4joKVGVB_5Vyph4g3ly1fiN4evdGtks5tQ2EpgaJpZM4QqzE->
.
|
|
Just to clarify where we are at, you can do all of those things now with xgcm, just excluding the points on the face edges. |
|
Jepp. I know. Thanks!
Global maps are pretty great though!
Particularly, the streamfunction on a face is not as exciting as the global
picture!
…On Fri, Feb 2, 2018 at 2:33 PM, Ryan Abernathey ***@***.***> wrote:
Just to clarify where we are at, you can do all of those things *now*
with xgcm, just excluding the points on the face edges.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#88 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AEY4jpkXLsO4utoICLJpJWUbS4sXUEQKks5tQ2MKgaJpZM4QqzE->
.
|
|
Despite the broken tests (related to grid autogeneration), I would like to merge this PR in order to keep things moving forward. In particular, I know that @maikejulie and @ifenty are eager to try it out with the ECCOv4 tutorials. I will fix the autogeneration stuff in a future PR. Would anyone like to provide a review before I merge? |
rabernat
changed the title
This is a first draft at modifying xgcm's guts to allow for more complex grid geometries, such as the MITgcm LLC grids, cubed-sphere grids, etc.
cc @maikejulie, @jetesdal