Re: [ferret_users] composite of active phase

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

Hi Jaya,
Yes, this method looks fine. The sampling that SAMPLEL does is based on 
index values that refer to L=1,2,3,....  so if you have listed the index 
values that correspond to the times you want to include, then averaging 
the sampled variable is the average over just those times.

To make a list of values before and after the chosen time points, I have 
an idea.  I will show an example to make a list of the index values 5 
points before and after a particular list of index values,and you can do 
the same thing with a list of 20 points before and after.  We will 
define a t variable from the list of indices, and an x variable centered 
on zero, with 5 points on each side. Then we'll use these variables to 
define a 2D variable whose values in the x direction will be the set of 
chosen index values, plus or minus 5.  Then we will create a 1D list of 
this new variable.

! A set of points which correspond toyour active phase dates, on a T axis
yes? LET tpts = TSEQUENCE({100,150,300}) 

! An axis in another direction, centered on zero
yes? DEFINE AXIS/X=-5:5:1 x5
yes? let xpts = x[gx=x5]

yes? LET xtvar = xpts + tpts

yes? list xtvar
            VARIABLE : X[GX=X5] + TPTS
            SUBSET   : 11 by 3 points (X-T)
           -5     -4     -3     -2     -1      0      1      2      
3      4      5   
            1      2      3      4      5      6      7      8      
9     10     11
1   / 1:   95.0   96.0   97.0   98.0   99.0  100.0  101.0  102.0  
103.0  104.0  105.0
2   / 2:  145.0  146.0  147.0  148.0  149.0  150.0  151.0  152.0  
153.0  154.0  155.0
3   / 3:  295.0  296.0  297.0  298.0  299.0  300.0  301.0  302.0  
303.0  304.0  305.0
                                               ^
                                               ^
          |---- 5 on each side ------ the original points ------ 5 on 
each side ----|

yes?  ! now make this into a 1D variable

yes? LET xtseries = XSEQUENCE(xtvar)
yes? list xtseries
            VARIABLE : XSEQUENCE (XTVAR)
            SUBSET   : 33 points (X)
1    /  1:   95.0
2    /  2:   96.0
3    /  3:   97.0
4    /  4:   98.0
5    /  5:   99.0
6    /  6:  100.0  <--
7    /  7:  101.0
8    /  8:  102.0
9    /  9:  103.0
10   / 10:  104.0
11   / 11:  105.0
12   / 12:  145.0
13   / 13:  146.0
14   / 14:  147.0
15   / 15:  148.0
16   / 16:  149.0
17   / 17:  150.0  <--
18   / 18:  151.0
19   / 19:  152.0
20   / 20:  153.0
21   / 21:  154.0
22   / 22:  155.0
23   / 23:  295.0
24   / 24:  296.0
25   / 25:  297.0
26   / 26:  298.0
27   / 27:  299.0
28   / 28:  300.0  <--
29   / 29:  301.0
30   / 30:  302.0
31   / 31:  303.0
32   / 32:  304.0
33   / 33:  305.0

See how the points in this list are the set of L indices originally 
chosen, plus 5 points on each side of them.  You can use this in a call 
to SAMPLEL.

yes? let var5  =SAMPLEL(var,xtseries)

If you follow this example, but define the x axis with /X=-20:20:1 I 
think you will have what you need.   The output of the samplel function 
is on an abstract axis, so the result of the example I just did, var5 
would have a T axis with T=1,2,3, ... 33.

Ansley

jaya kumar A wrote:
> dear ferret users,
> I want to calculate the composite of active phase which corresponds to 
> different dates in the JJAS ,
> if i pick the dates from the different years I tried like this
> let SD_posv=SAMPLEL(PRE_ANOM_NEW[d=1],{171, 204, 195, 2002, 12.....})
> con/lev=(0.2,2,0.2)/nolab/nokey SD_posv[l=@ave]
>
> will this method is ok?
> if it is wrong any good method to make
> I also want to make a time series of 20 days before and after active 
> phase so  if i define SD_posv as active phase (above)
>
> with regards
> jaya
>
> --
> JAYAKUMAR A
> Senior Research Fellow,
> Theoretical Studies Division
> IITM, Pune-8