[ecco-support] gcmfaces_lines_transp and calc_trasnports

[gael forget]

Dear Paola,

it seems that the answer I sent last week bounced somehow so I copy and paste it here:
> Hi Paola,
> I suspect that the grid line path generated by gcmfaces_lines_transp_lat.m is affected by the change in end points. If that is the case you should see differences mskCedge etc.
> Hope this helps,
> Gael

ps: thanks Ian for pointing out that my email had bounced

> On Feb 23, 2018, at 10:08 PM, Fenty, Ian G (329C) <Ian.Fenty at jpl.nasa.gov> wrote:
> 
> Dear Paola, 
>  
> Land points in ECCO v4 are defined on a curvilinear grid whose parameters are laid out in the different files of the nctiles_grid/ directory.   The grid parameters for your domain are found in nctiles_grid/GRID.0012.nc.  Each model grid cell has a land mask defined at their center XC (longitude), YC (latitude).  Each cell has four corners, defined at the points XG, YG. 
>  
> I plotted the land mask (hFacC) below.  Land points are empty white boxes and water points have red dots.  The grid cell boundaries are plotting using XG and YG.   
>  
> There is a water grid cell that spans approximately 72W to 71W at 30S.  The grid cell to the east of that cell spans approximately 71W to 70W and is land.  At first glance it would appear that the transport should not change when you move the eastern endpoint to 70W from 71W.  However, you are picking points very close to the grid cell edge.  The curvilinear grid is not defined with exactly integer meridians so this might be the source of the problem. 
>  
> Can you please try to redo the calculation using 69W as the endpoint to see if you get the same value as when 70W is the endpoint?  If so, then it would suggest that the first calculation was not counting the grid cell between 72W and 71W in the transport calculation.  
>  
> Perhaps Gael might have more to say about this as he wrote the calc_transports routine.
>  
> Do let us know what you find.  
>  
> Ian
>  
>  
> <image001.png>
> On 2/22/18, 4:16 PM, "ecco-support-bounces at mit.edu <mailto:ecco-support-bounces at mit.edu> on behalf of Paola Cessi" <ecco-support-bounces at mit.edu <mailto:ecco-support-bounces at mit.edu> on behalf of pcessi at ucsd.edu <mailto:pcessi at ucsd.edu>> wrote:
>  
> I am confused as why the values of transports  across a section, as given by calc_transports, change when the endpoints of the section, as given by calc_transports, are changed by one or two degree on land points. For example, if I change gcmfaces_lines_transp([152 -71],[-30 -30],{'30S Pacific'}); to gcmfaces_lines_transp([152 -70],[-30 -30],{'30S Pacific'}); the transports are different even though <>
> both 70W 30S and 71W 30S are on land (at least on google maps). Why does this happen? Is there documentation on the actual land point in ECCO, given in lat-lon? 
>  
> Thanks,
> 
> Paola
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/pipermail/ecco-support/attachments/20180226/6eb52715/attachment.html
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 1839 bytes
Desc: not available
Url : http://mailman.mit.edu/pipermail/ecco-support/attachments/20180226/6eb52715/attachment.bin