Project

General

Profile

Help for 90 timpctl CDO

Added by Lili V almost 12 years ago

Hallo everyone,

I have a simple case.
I calculated 90 percentile from daily 1 deg precip dataset 1990-1995
for a bounded area. I did:

cdo -r timpctl,90 prep.nc -timmin prep.nc -timmax prep.nc prep99.nc

It was Ok, and i got my result in 2D (latxlon i.e. 99th percentile for each grid).
Actually, this is not what i am expecting. I want to have only 1 value of 99th percentile
for this area (not each grid). Averaging the 99th percentile in all grids is not correct. How could
i calculate the 99th percentila in a bounded area.

Please give me a hint and suggestion.
Thanks

Cheers,
Verena

p.s : result was attached.

90p.gif (36 KB) 90p.gif

Replies (3)

RE: Help for 90 timpctl CDO - Added by Jaison-Thomas Ambadan almost 12 years ago

try the field percentiles, "fldpctl" operator.

Cheers,
J.

RE: Help for 90 timpctl CDO - Added by Lili V almost 12 years ago

Hi J,

Thanks, but not that one.
I have daily 5 years datasets in a bounded area let say tropics.
I think what i mean is something like firstly
order all data from all grids to create
a histogram for the whole tropics to get 99th percentile (have only
one value).
Any suggestion how to do this in CDO?

Thanks

Cheers
Verana

RE: Help for 90 timpctl CDO - Added by Jaison-Thomas Ambadan almost 12 years ago

Hi,

As far as I know, there is no direct way to do this with CDO. Anyway, please have a look at the following thread, may help you:

https://code.zmaw.de/boards/2/topics/963#message-1025

Cheers,
J

    (1-3/3)