[Thread Prev][Thread Next][Index]

Re: [ferret_users] BUG REPORT: TSERIES creates spurious modulo axes



Hi Russ,

Thank you for the report.  We're just doing some work on coordinate axis handling to finally (!!)  get rid of the fixed maximum limit on coordinate storage and forever get rid of

 **TMAP ERR: limit on storage for coordinates has been reached   MAX=...

I will make sure that we address this.

By default Ferret has always defined axes as modulo if they start in year 0000 or 0001 and are a year or less in length. I haven't looked in detail, but the aggregation starts with the first dataset's time axis, and builds the aggregated axis using its attributes (checking that the time units are consistent, and also checking the time origin on subsequent axes and translating them if needed).  Clearly we're missing a final step of checking whether the aggregated axis should be modulo.


Ansley

On 9/5/2017 1:00 AM, Russ Fiedler wrote:

Hi,


I'm aggregating some files from MOM output data with TSERIES but Ferret insists on making the time axis "modulo" for some reason.

Issuing "CAN AXIS/MOD time"  fixes things but it's a pain.

Each file has 12 months of data

From an ncdump on a file

...

    time = UNLIMITED ; // (12 currently)
...

    double time(time) ;
        time:long_name = "time" ;
        time:units = "days since 0001-01-01 00:00:00" ;
        time:cartesian_axis = "T" ;
        time:calendar_type = "NOLEAP" ;
        time:calendar = "NOLEAP" ;
        time:bounds = "time_bounds" ;


There's no modulo attribute here.

But SHOW AXIS

 time      TIME                24mi   16-JAN 12:00 16-DEC-0002 12:00         <---- start time has no year!

It looks to be a problem when the base year is 0001.

When I changed the base year in the units string to 0010 it looks fine

 time      TIME                24 i   16-JAN-0010 12:00 16-DEC-0011 12:00


On further investigation it looks like a general problem with having a file with data only in the year 0001. Ferret seems to get a bit too frisky and wants to make climatologies no matter what.


If I load only the first year a climatology is automatically made

 time      TIME                12mi   16-JAN 12:00         16-DEC 12:00

if I load the second year by itself it's fine

 time      TIME                12 i   16-JAN-0002 12:00 16-DEC-0002 12:00


If I concatenate the two files using ncrcat no climatology is made.

 time      TIME                24 i   16-JAN-0001 12:00 16-DEC-0002 12:00

Cheers,

Russ






[Thread Prev][Thread Next][Index]
Contact Us
Dept of Commerce / NOAA / OAR / PMEL / Ferret

Privacy Policy | Disclaimer | Accessibility Statement