Re: leap year and mc files

[Brent A McDaniel <gte328r@xxxxxxxxxxxxxxxxxxxxx>]

Ansley,

I've upgraded to a version that sees the noleap calander properly when I 
specify time time:calendar = "noleap" ; in the netcdf file.  By this I 
mean that if I list the feb-mar time series the data goes from feb28 to 
mar 1 (correctly).  I'm still struggling with the .des file though.  I've 
included a short .des file to show you what I'm using.  


The problem is the s_start/ s_end for the 1980 data file.  If I specify 
17347584-17356344 I get an error when I try to read the value that should 
correspond to dec/31 (ferret is seeing this as the 366th day of the year, 
where's there's no data in the file, so error).  If I use 
17347584-17356320, there's no error in the time series but it stops at dec 
30, 1980 (the 365 day of the time series, the last data value in the 
file).  Either way, it's reporting a value for feb 29 even though 
nascar.1980.nc.1980 has calender = "noleap".  

Not sure if this matters but all the files have the time origin of their 
own year.  IE, for 1979, ncdump -h shows time:units = "days since 
1979-01-01 12:00:00.0" ;

Any help's appreciated, I'm stuck

Cheers,

Brent




test.des

&FORMAT_RECORD
  D_TYPE               = '  MC',
  D_FORMAT             = '  1A',
  D_SOURCE_CLASS       = 'MODEL OUTPUT',  
/
&BACKGROUND_RECORD
  D_EXPNUM             = '0000',
  D_MODNUM             = '  00',
  D_TITLE              = 'Model Output, Daily Averages',
  D_T0TIME             = '30-DEC-0000 00:00:00',
  D_TIME_UNIT          = 3600.,
  D_TIME_MODULO        = .FALSE.,
/
&MESSAGE_RECORD
  D_MESSAGE            = ' ',
  D_ALERT_ON_OPEN      = F,
  D_ALERT_ON_OUTPUT    = F,
/
&EXTRA_RECORD
/

&STEPFILE_RECORD
  S_FILENAME         = "/data/nascar/nascar.1979.nc.1979",
  S_AUX_SET_NUM      = 0,
  S_START            = 17338824.,
  S_END              = 17347560.,
  S_DELTA            = 24.,
  S_NUM_OF_FILES     = 1,
/

&STEPFILE_RECORD
  S_FILENAME         = "/data/nascar/nascar.1980.nc.1980",
  S_AUX_SET_NUM      = 0,
  S_START            = 17347584.,
  S_END              = 17356344.,
  S_DELTA            = 24.,
  S_NUM_OF_FILES     = 1,
/


&STEPFILE_RECORD
  s_filename           = '**END OF STEPFILES**'
/













On Mon, 8 Apr 2002, Ansley Manke wrote:

> Hi Brent,
> Do upgrade to version 5.4.   Ferret v5.22 doesn't recognize the "noleap"
> attribute.   I think that the using a descriptor file will work fine if you
> specify the calendar attribute in the annual files and use the actual start
> and end dates in the .des file.   Let us know how that goes.
> 
> Ansley Manke
> 
> Brent A McDaniel wrote:
> 
> > Hi Ferret Users,
> >
> > I have begun analysis on data produced by the NASCAR model which defines
> > every year as 365 days (so no leap years on 1980 etc).  I converted the
> > grads data to netcdf via grads2nc and opened it with ferret.  No suprise,
> > it assumed that 1980 was supposed to have leap years so the dates were off
> > by one.  I remade the netcdf file using the
> > time:calendar_type = "noleap" ; in the header and opened the file via
> > ferret and still it assumed that 1980 was a leap year (putting the data
> > that should have been mar1 at mar0 and skewing all later data by 1 day).
> > Is there a good way to get ferret to recognize the data as being of the
> > type noleapyear?
> >
> > I'm still using v5.22 but if I upgraded, I could define a new noleap axis
> > and regrid the data to it but this seems clunky and is less than
> > desireable.  Any suggestions?
> >
> > Second question:  Once I figure out question 1, I want to open 17
> > annual data files as 1 via the mc descriptor files.  I have .des files
> > that I use for standard calenders, but I'm confused as to what to put in
> > the s_start and s_end because of the lack of leap years.  Is there a flag
> > I can somehow use to tell ferret that I want to use noleap year calenders?
> > If someone has an example of a noleap year .des file that would be great
> > as well.
> >
> > Hopefully I've explained myself well enough to make my troubles clear.
> >
> > Thanks as always for the help.
> >
> > --Brent
> >
> > --
> > Brent A. McDaniel
> >
> > Dept of Earth and Atmospheric Sciences
> > Georgia Institute of Technology
> > Atlanta, Ga.  USA
> 
> --
> Ansley Manke  Pacific Marine Environmental Laboratory  Seattle WA  (206)526-6246
> 
> 

-- 
Brent A. McDaniel

Dept of Earth and Atmospheric Sciences
Georgia Institute of Technology
Atlanta, Ga.  USA