Re: [ferret_users] Composite of MJO phases

Dear Rahmi,

Maybe you can make a composite (average cases) with netcdf operators like NCO or CDO.

I mean you can select the cases of any MJO phase and make a average. In cdo you could you operator -seltime to select the events.

Further, for significance level of your composite you should calculate the effective number (equivalent) of your sample, Because could be autocorrelation in the data, then the size of your sample could be smaller. An estimation of the sample size could be done with http://clivac.geog.ucsb.edu/wp-content/uploads/dependence_files_image006.gif
where n is the size of sample, p1 is autocorrelation with lag-1 and n' is the effective number of independents events for composite.

You will use the n' for estimate the degrees of freedom (v), you could be and look up the t-student table to find the critical t-value.

For significance level in composites you can check here:
http://clivac.geog.ucsb.edu/geog-288-lc/significance-tests/

and about the autocorrelation here
http://clivac.geog.ucsb.edu/geog-288-lc/autocorrelation/

I hope that helps,

Cheers,

--
------------------------------------------------------------
Francisco das Chagas Vasconcelos Junior
PhD Candidate
Dept. of Atmopheric Sciences
University of Sao Paulo - USP
Institute of Astronomy, Geophysics and Atmospheric Sciences - IAG

On Wed, Nov 27, 2013 at 11:10 PM, rahmi ariani <rahmi.ariani@xxxxxxxxx> wrote:

Dear Ferret users,
I have 11 years netCDF data which is the output of Regcm4 Model. I want to make a composite of precipitation data that have the same MJO phases from that file. Let say MJO phase 1 occurred on (yyyy-mm-dd):
20000926
20000927
20001110
20001111
20001112
20021211
20021212
20021213
...
...
20100115
20100116
...
...

I put the list on a .txt file. So, in the end i will have 8 netCDF file that content the same MJO phases on 11 years. Any advice to do this composite? Thank you in advance.

Here is the description of my data (RegCM4 output):
    XAXIS = 103 ;
    YAXIS = 95 ;
    TIME = UNLIMITED ; // (2928 currently)
    bnds = 2 ;
variables:
    double XAXIS(XAXIS) ;
        XAXIS:units = "DEGREE" ;
        XAXIS:point_spacing = "even" ;
        XAXIS:axis = "X" ;
        XAXIS:modulo = 360. ;
    double YAXIS(YAXIS) ;
        YAXIS:units = "DEGREE" ;
        YAXIS:point_spacing = "even" ;
        YAXIS:axis = "Y" ;
    double TIME(TIME) ;
        TIME:long_name = "time" ;
        TIME:calendar = "gregorian" ;
        TIME:units = "hours since 1949-12-01 00:00:00" ;
        TIME:bounds = "TIME_bnds" ;
        TIME:axis = "T" ;
    double TIME_bnds(TIME, bnds) ;
    float PR(TIME, YAXIS, XAXIS) ;
        PR:missing_value = -1.e+34f ;
        PR:_FillValue = -1.e+34f ;
        PR:long_name = "PRCV[GXY=NEWGRID@ASN]" ;
        PR:history = "From MIT_20_Sumatera_SRF.2000010100" ;

Kind regards,

--
Rahmi Ariani
Graduate Student of Applied Climatology
Faculty of Mathematics and Natural Science
Bogor Agricultural University
Indonesia
16680