ttest......

Hello support

I want to do the ttest of composite anomaly. My null hypothesis is composite
anomaly is not statistically significant. . The variable "JJAS_olr_anom_E"
is having dimension lat=73,lon=144,time=4. The variable
"JJAS_olr_anom_1983_05" is having dimension lat=73,lon=144,time=23. aveX,
varX, aveY,varY is having dimension lat=73,lon=144
so my sX=4, sY=23. I want to plot only statistically significant value only.
Anybody can help me?

Any help would be appreciated. I am attaching the script here

load "$NCARG_ROOT/lib/ncarg/nclscripts/csm/gsn_code.ncl"
load "$NCARG_ROOT/lib/ncarg/nclscripts/csm/gsn_csm.ncl"
load "$NCARG_ROOT/lib/ncarg/nclscripts/csm/contributed.ncl"
begin
;**************************************************************
in2=addfile("mn_olr_1983-05.nc","r")
olr=in2->olr
olrflt=short2flt(olr)
mn_clim_olr=clmMonTLL(olrflt)
;printVarSummary(mn_clim_olr)
Mon_olr_anom=calcMonAnomTLL(olrflt,mn_clim_olr)
Mon_olr_anom@_FillValue=32767
Mon_olr_anom_at_missing_value=32767
;*****************************************************
; significant test
;*********************************************************************
four_month_runavg=runave_Wrap(Mon_olr_anom(lat|:,lon|:,time|:),4,0)
JJAS_olr_anom_1983_05=four_month_runavg(:,:,6::12)
ii=(/0,7,16,21/)
JJAS_olr_anom_E = (/JJAS_olr_anom_1983_05(:,:,ii)/)

siglvl=0.05
aveX=dim_avg_n(JJAS_olr_anom_E,2)
varX=dim_variance_n(JJAS_olr_anom_E,2)
sX=4
sY=23
aveY=dim_avg_Wrap(JJAS_olr_anom_1983_05)
varY=dim_variance_Wrap(JJAS_olr_anom_1983_05)
iflag=True
probt = ttest(aveX,varX,sX, aveY,varY,sY, iflag,False)
aveX=mask(aveX,probt.lt.siglvl,True)
copy_VarCoords(olr(0,:,:),aveX)
;***********************************************************************
wks=gsn_open_wks("ps","test_composite")
gsn_define_colormap(wks,"BlueRed")
res=True
res_at_cnFillOn=True
res_at_gsnSpreadColors=True
res_at_cnLinesOn=False
plot=gsn_csm_contour_map_ce(wks,aveX,res)
;***********************************************************************
end

14,1 Top

_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Thu Sep 03 2009 - 06:26:39 MDT