Era-Interim masking help

[Karl]

Hi I have two .nc files (a land-sea mask and rainfall both for era-interim).

I am trying to calculate the time-series over a grid-box, however this grid-box includes the sea values as well which i do not want. To exclude the sea values i am trying to mask off all values over the sea, and have some difficulties in this aspect.

What i've done so far:
1) I removed a degenerate time dimension from the land-sea mask (the land-sea mask fron ERA-Interim only contains one variable 'lsm').
2) Appended the land-sea mask into my rainfall data

I've also ensure the latitude, longitude size of both files are the same, as well as their resolution.

From here I am not sure of what i should do, in previous threads that i have looked, some have suggested using ncwa or ncap2. Any help and advice is appreciated!

[henry Butowsky]

HI Karl,
you are almost there --
ill assume that rainfal var is "double pcp(time,lat,lon) " and that it has a missing_value or _FillValue
and the land-sea mask is "int mask(lat,lon)" ( 1 is land , 0 is sea )

/******* avg.nco*********/
  where( mask ==1 )
      pcp=pcp;
    elsewhere
      pcp=pcp.get_miss();

     pcp_avg=pcp.avg($lat,$lon); 
  /****************************/    

to run the above script us ethe following command.

ncap2 -v -O -S avg.nco in.nc out.nc

...Henry

[Karl]

Hi Henry, thank you for your reply.

A quick ncdump on my rainfall data gave me:

    short tp(time, latitude, longitude) ;
            tp:scale_factor = 5.94397276467581e-06 ;
            tp:add_offset = 0.194760211607368 ;
            tp:_FillValue = -32767s ;
            tp:missing_value = -32767s ;
            tp:units = "m" ;
            tp:long_name = "Total precipitation" ;
    double lsm(latitude, longitude) ;
            lsm:_FillValue = -32767. ;
            lsm:missing_value = -32767s ;
            lsm:units = "(0 - 1)" ;
            lsm:long_name = "Land-sea mask" ;
            lsm:standard_name = "land_binary_mask" ;

Would the above script still be applicable?

Cheers

[henry Butowsky]

HI Karl,
yep should still work -
just change lat to latitude and lon to longitude.

...Henry

[Karl]

Hi Henry,

i am receiving their error mssg after running ncaps2 line:
avg.nco:2:15: expecting (, found '=='

I am not sure what there error means..

Karl

[henry Butowsky]

sorry mask is a function -use yor var name s in the script e.g

/******* avg.nco*********/
  where( lsm ==1 )
      tp=tp;
    elsewhere
      tp=tp.get_miss();

     tp_avg=tp.avg($latitude,$longitude); 
  /****************************/