velocity, salinity, and temperature on 5/4/2012

[Lianyuan Zheng]

Hello,

The HYCOM output velocity, salinity, and temperature on 5/4/2012 are
incorrect. Please check it. Thanks.

Lianyuan

[Michael McDonald]

Lianyuan,

I assume you are talking about the GLBa0.08 model, expt_90.9 data
(*not specified in your email*).

Catalog URL: http://tds.hycom.org/thredds/GLBa0.08/expt_90.9.html?dataset=GLBa0.08/expt_90.9
OPENDAP URL: http://tds.hycom.org/thredds/dodsC/GLBa0.08/expt_90.9

Can you elaborate more on the variables "incorrectness"?

Were you receiving NaN for all areas, or were you actually getting
incorrect data (i.e., large jumps in values between days)?

If you were trying to acces the data between these times (EST),

2012-04-30 18:58:43 Started
[ 18:58:43 ] <<EXTRACTION STARTED>>
[ 19:51:17 ] <<PROCESSING STARTED>>
[ 00:25:40 ] <<PUBLISHING STARTED>>
[ 01:31:51 ] <<COMPRESSION STARTED>>
2012-05-01 03:14:17 Completed

Then you might have been querying the latest "04-MAY-2012" data while
the new files were being written to the data server. The "publishing"
of the data overwrites some hindcast data (3-5 day overlap window) and
all forecast data with new netCDF files nightly. Do you have a more
precise timeframe when your data requests failed?

/mike

[Lianyuan Zheng]

Mike,

I downloaded data from the Aggregated OpenDap: http://tds.hycom.org/thredds/dodsC/glb_analysis
every morning around 1:00am. All the velocity, salinity, and
temperature on the last time step (i.e. 5/4/2012) are either 0 or
missing_value (1.E+39) (Note: all data before 5/3/2012 are normal).
Before I send this email out, I check the OpenDap file again and it is
correct now. Have you do anything on the Aggregated output file
between 9:30am and 1:45am?

Thanks,
Lianyuan

On May 1, 1:49 pm, Michael McDonald <michael.mcdon...@hycom.org>
wrote:

> Lianyuan,
>
> I assume you are talking about the GLBa0.08 model, expt_90.9 data
> (*not specified in your email*).
>

> Catalog URL:http://tds.hycom.org/thredds/GLBa0.08/expt_90.9.html?dataset=GLBa0.08...

> OPENDAP URL:http://tds.hycom.org/thredds/dodsC/GLBa0.08/expt_90.9
>
> Can you elaborate more on the variables "incorrectness"?
>
> Were you receiving NaN for all areas, or were you actually getting
> incorrect data (i.e., large jumps in values between days)?
>
> If you were trying to acces the data between these times (EST),
>
> 2012-04-30 18:58:43 Started
> [ 18:58:43 ] <<EXTRACTION STARTED>>
> [ 19:51:17 ] <<PROCESSING STARTED>>
> [ 00:25:40 ] <<PUBLISHING STARTED>>
> [ 01:31:51 ] <<COMPRESSION STARTED>>
> 2012-05-01 03:14:17 Completed
>
> Then you might have been querying the latest "04-MAY-2012" data while
> the new files were being written to the data server. The "publishing"
> of the data overwrites some hindcast data (3-5 day overlap window) and
> all forecast data with new netCDF files nightly. Do you have a more
> precise timeframe when your data requests failed?
>
> /mike
>

[Michael McDonald]

> I downloaded data from the Aggregated OpenDap: http://tds.hycom.org/thredds/dodsC/glb_analysis
> every morning around 1:00am.  All the velocity, salinity, and
> temperature on the last time step (i.e. 5/4/2012) are either 0 or
> missing_value (1.E+39) (Note: all data before 5/3/2012 are normal).

What application(s)/method(s) are you using to download the data?

> Before I send this email out, I check the OpenDap file again and it is
> correct now.  Have you do anything on the Aggregated output file
> between 9:30am and 1:45am?

Yes. The server is updated with new data files nightly
(post-processing of hycom a/b to netCDF) starts at 8pm EST and
finishes around 8am EST. During this time there are only brief windows
(~20 seconds / file) where each of the latest data files are being
overwritten with the best hindcast data and the latest forecast data.
You were just unlucky enough to be reading this data at that precise
time.

The restarting of the tomcat server at Noon (EST) daily is our
mechanism to rescan/update the new forecast data that has been added
that evening/morning.