Walt,
Sigh. Yes, this is a known problem with NetCDF-java/IDV and staggered
grids. Plotting 3D velocities with IDV will not work with output from
staggered grid models like ROMS (and the ocean model component of
COAWST is ROMS).
The problem is that the U and V velocities are at the grid cell walls,
while the vertical coordinate is defined on the grid cell center. The
result is that NetCDF-Java returns an 3D array for Z that is one
column wider than the U values (or one row wider for the V values).
A simple solution, which we hope to implement and contribute back to
NetCDF-Java, would be to check the size of the returned arrays, and if
they don't match, average to the U and V points.
This only applies to the 3D velocities. You can plot the 2D
depth-averaged velocities in IDV, (although when I plotted the COAWST
depth-averaged velocities as "grid-relative u and v components", they
don't look correct -- I'll check into this).
Thanks,
Rich
On Wed, Jun 2, 2010 at 6:25 PM, Walt McCall <Walt.McCall@xxxxxxxx> wrote:
> Rich,
>
> Tried loading the COAWST->COAWST Nowcast->coawst 2->forecast Model Run model
> under GCOOS from your thredds server and was not able to load the U, V, or
> Flow Vectors. Might have a matrix mismatch.
>
> All the other fields seem to work OK.
>
> Anybody else run into this problem?
>
>
> Couldn't get data SampledSet.init_samples: sample#2 length 4816896 doesn't
> match expected length 4811520
>
> Thanks,
> Walt
>
> Rich Signell wrote:
>>
>> IDV Enthusiasts,
>>
>> For those of you with an interest in the circulation models being used
>> for oil spill forecasts, here is an IDV bundle showing surface current
>> and salinity (since it's more conservative than temperature) from two
>> different ocean forecast models.
>>
>> http://coast-enviro.er.usgs.gov/models/share/gulf_salt_currents2.xidv
>> (1. Save to disk and then "File=>Open", or 2. "Data=>Choose data=>From
>> a Web server" and paste in the above URL)
>>
>> You need to be a bit patient, as it takes about 3-4 minutes of
>> downloading data (on my connection anyway) before you see anything.
>> Once it's loaded, you can animate, and switch between the two models.
>> Taking advantage of Jeff's sneaky "reverse times" option, it will
>> always display the last 3 days from each model if you reload data.
>> This also means one model may have an additional day on the front or
>> end of the other model, so if you see no data with a particular model
>> selected, click to the next or previous time.
>>
>> Right now one of the models is showing an attached loop current, the
>> other a detached ring!
>>
>> More models of the Gulf can be accessed via this catalog (scroll down
>> to the GCOOS category):
>> Data=>From Catalog=>http://coast-enviro.er.usgs.gov/thredds/ioos_clean.xml
>>
>> Let me know if you have success, failure, or suggestions, or if you
>> come up with better bundles!
>>
>> I'm working on IDV 2.9, so if something doesn't work try upgrading IDV
>> first.
>>
>> -Rich
>>
>
--
Dr. Richard P. Signell (508) 457-2229
USGS, 384 Woods Hole Rd.
Woods Hole, MA 02543-1598
