search

NOAA-PMEL / LAS

Live Access Server
https://ferret.pmel.noaa.gov/LAS/
The Unlicense
13 stars 5 forks source link

Curvi-linear plot fails in "Compare 2" with COADS in the left panel. #1741

Closed karlmsmith closed 6 years ago

karlmsmith commented 6 years ago

Reported by @noaaroland on 14 Jun 2016 18:48 UTC Using this tri-polar data set:

OPeNDAP URL: http://data1.gfdl.noaa.gov:8080/thredds/dodsC/CM2.1R_ECDA_v3.1_1960.r1i1p1.dc_Omon

It plots great on it's native grid, but when used in a "Compare 2" with the default COADS (which is on a slightly off-set longitude range) the resulting plot is messed up.

Migrated-From: http://dunkel.pmel.noaa.gov/trac/las/ticket/1732

karlmsmith commented 6 years ago

Attachment from @noaaroland on 14 Jun 2016 18:50 UTC Messed up plot. a8f3583a09d8f1da63821e0ae6d8e28b_plot_image

karlmsmith commented 6 years ago

Attachment from @noaaroland on 14 Jun 2016 18:56 UTC The state of the UI. ui

karlmsmith commented 6 years ago

Comment by @AnsleyManke on 14 Jun 2016 20:00 UTC This can be made to work - the plot script needs to apply and HLIMIT on the plot command.

karlmsmith commented 6 years ago

Comment by @AnsleyManke on 15 Jun 2016 00:27 UTC The plot looks so odd because the scripts are matching the XY region with the plot on the left. The logic to determine the index ranges in a curvilinear grid that correspond to any XY region wasn't robust enough and does funny things with these particular curvilinear grids.

I am backing off that method, making the plot with the whole region, limited just by the plot axes. This is not ideal. Too much data will be read.

When there is a pre-computed map file that will still be used to compute the index region, and it should be reliable. This is yet another reason to pre-computed mappings for GFDL grid(s), as well as what's in Ticket 1733.

I have checked in the fix on the las.v8.4.rc4 tag.