Re: [ferret_users] strange error in climatological zonal average

[Ansley Manke <ansley.b.manke@xxxxxxxx>]

Hi,
The order of operations should not make a difference here.  I found a 4-D dataset and tried your commands and was not able to reproduce the no-valid-data behavior.

What happens if you remove the SET REGION command and instead apply the region when you do the averaging, that is,
fill/nolab ztempc[x=40:110@ave,z=-10:40@ave]

Another thing to do in finding out what is happening as a script runs, is to turn on the diagnostic mode, "SET MODE DIAGNOSTIC".  There's an FAQ about interpreting the output,
How do I interpret the output of "SET MODE DIAGNOSTIC?"

It tells you as the commands are being processed, what index ranges are being used for creating variables.  If you compare the case that is not working with a similar one which is, you might be able to spot what the problem is.

Ansley

On 11/22/2011 8:40 AM, jagadish karmacharya wrote:

Hi all,

I am facing a trouble in zonal averaging that seems strange, atleast to me. The error seems to occur only when condition is just right (or shall I say wrong). Here is my script:

use "/data/erai/erai.ZT.mon.abs.nc"
set region/x=40:110/y=-10:40
set data climatological_axes; can data climatological_axes
let zta=ztemp[t=15-dec-1991:15-nov-2005]
let ztempc=zta[gt=month_reg@mod]
fill/nolab ztempc[x=@ave,z=@ave]

This results in " No valid data " in the plot window. Here is details about data:
sh d
    currently set data sets:
    1> /data/erai/erai.ZT.mon.abs.nc (default)
name        title                        I            J         K        L
ZTEMP    Temperature           1:49     1:38      1:8      1:269   

I tried to trace the source of error without sucess but found that  the error doesn't occur under different scenarios such as:
1) Meridional average ( i.e. fill/nolab ztempc[y=@ave,z=@ave] -> generate plot)
2) use whole time seris used to produce climatology (i.e. let ztempc=ztemp[gt=month_reg@mod]  & then using fill -> generate plot)
3) if zonal average is done ahead of fill command (i.e  replace one of the corresponding line in above script with
    let zta=ztemp[x=@ave,t=15-dec-1991:15-nov-2005]
    or 
    let ztempc=zta[x=@ave,gt=month_reg@mod]
    and then fill/nolab ztempc[z=@ave] -> generate plot)
4) no such problem with one other data I tried:
    (sh d
         currently set data sets:
         1> /data/model/aktica.pm.ztemp.nc (default)
         name        title                        I             J         K        L
        ZTEMP    Temperature T          1:159   1:113    1:17    1:228 )

I am wondering why this error occured in the 1st scenario? does it has to do with data order or something else?

Thank in advice.
 

Jagadish