Martin, Ryo,
thanks for your answers. Actually @ave is doing weight averaging for vertical averaging. I know that it will also do the same for horizontal averaging. Unfortunately, in my file (please find attached) latitude and longitude are not aligned with x,y axis
so is I do:
yes? stat/k=1 temp
...
Mean value: -1.7884 (unweighted average)
and using @ave
yes? list/k=1 temp[x=@ave,y=@ave]
...
-1.788
but I'd like to have area weighed averaged value. How should I told ferret that x and y is long and lat for averaging computations,
thanks,
Robert
From: Martin Schmidt [martin.schmidt@xxxxxxxxxxxxxxxxx]
Sent: Thursday, March 02, 2017 7:20 AM To: Osinski, Robert FORNATL, PL; ferret_users@xxxxxxxx Subject: Re: [ferret_users] vertical weighted averagng Hi Robert,
a first inspection with ncdump shows an overly long file history. There is a flag in nco to supress this. I have the feeling the file is aggregated from time slice files using nco. This is not needed if you are working locally, since ferret can do the aggregation for you. Please consult the manual. For sure, it is appropriate to select one grid point, to send the file via mail. The first step should be to read the file header with ferret ferret use sample_TS.nc show data You may also want to see the vertical distribution of grid points: list z[gz=temp] To average temp, you may use the ferret operator @ave list/l=1 temp[z=@ave] You may plot a time series plot temp[z=@ave] This looks simple, but for sure ferret makes assumptions where exactly the grid points are positioned within the grid cells. Any information on this like on cell boundaries is missing in the file. So this needs some thoughts. So far, the result should be fine. Hope this gets you started. I suggest, to have a look into the excellent manual and go through the tutorial. Best, Martin On 03/02/2017 02:56 PM, Osinski, Robert FORNATL, PL wrote:
|
Attachment:
sample_file_POP.nc
Description: sample_file_POP.nc