McIDAS-V Support Forums

On November 17th, a McIDAS-V Brainstorming session was held for all scientists at SSEC and CIMSS. At the session, one basic question was asked of the users: What would you like to see in McIDAS-V?

The following discussion is from that brainstorming session:

There are still gaps at dateline when displaying global data on flat 2d projections. (RickK)

* Rick said that although progress has been made, he still sees gaps at the dateline when displaying global data on flat 2d projections. It looks fine (no gaps at dateline) when projected on the 3d globe, but the gap appears when it's changed to a flat 2d projection. This can be a real problem in the SOS (science on a sphere) project tasks.

I wrote up inquiry #943 for this - http://mcidas.ssec.wisc.edu/inquiry-v/?inquiry=943

Is there a way to display GFS data across the dateline other than displaying the all data west of the dateline and then adding a second layer with all data east of the dateline?

Thanks
Joleen

Hi Joleen -

I just wanted to get some clarification on what you are asking about. Are you trying to subset in the Region tab across the dateline where the left/right bounds are at the dateline? For example, with GFS 2.5 degree global data across the SSEC catalog, I see the following in the Region tab:

In this case, it's easy to subset across the dateline, but not the prime meridian (0 degrees longitude).

Thanks -
Bob Carp

Hi Bob,

Sorry, I misdirected you. I meant the prime meridian.

Joleen

Thanks for the clarification, Joleen. This issue is written up as Inquiry 1626. Looking up the issue online, I saw DonM at Unidata mentioned to a user:

The problem is that these datasets go from 0-360 and we do not have a way in the underlying netCDF-Java layer to create a subset across the grid boundary.

See: https://www.unidata.ucar.edu/support/he ... 02671.html

It looks like this is related to THREDDS improvement #259, which is listed as being a 'major' priority.

I haven't tried this, but there was a user who found a workaround by converting a local grib2 file to a grib1 file using some extra software (Cnvgrib and Convsh + Subset.tcl).
See: http://www.unidata.ucar.edu/mailing_lis ... 00017.html

I added a note to the inquiry to keep you updated if there is any progress on this.

Thanks -
Bob

I would like to know if there has been any progress on this issue. When I try to visualize the GFS outputs over Europe (projections->Europe) and use Match Display Region, only data east of meridian 0 appears.

Thanks

Hello,

There are no updates to the status of this problem and I verified that it's still around in our nightly build. I tried changing the units attribute of the longitude variable in the GFS global data from "degrees_east" to "degrees_west" but this didn't help too much, the data only drew to the left of the prime meridian after this change.

One thing I noticed with the Europe projection is that once you zoom out of the display, a large portion of the data is actually loaded and displayed, I'd estimate that if the data displayed on both sides of the dateline that almost 1/2 of the global contours would be drawn. You're likely already aware of this, but a workaround for now would be to:

1. Change the the Europe projection in your display.
2. Disable "Auto-set Projection" with the "Projections > Auto-set Projection" menu item in the Main Display. This will keep the display in the Europe projection instead of snapping to the native projection of the data at Create Display.
3. Display the data using "Use Default" to display contours over the entire globe.

This will draw more contours than necessary if everything was working expected, but it might not be too much of a performance hit.

I'll let you know if there is any progress on this bug, but as the inquiry notes it's likely a lower-level netCDF issue that we might not have much control over.

Thanks,
Bob Carp
McIDAS Help Desk

Thank you so much, Bob.

Ricardo