为什么变化最大的正负闭合中心通不过显著性检验

我是小猪 · 发表于 2017-4-4 23:26:20

登录后查看更多精彩内容~

您需要登录才可以下载或查看，没有帐号？立即注册

x

如图，在热带地区以外的最值闭合中心却通不过显著性检验，反而旁边的区域却通过了，感觉像是检验填色图左移几个格子似的，不知道怎么回事？望各位朋友帮忙解决一下，在此谢谢各位啦！

（代码较长，有些地方略啰唆，代码为：300hPa位势高度与EOF的PC1做回归，并对其做t检验（tval））
十分感谢各位留言的童鞋！！！

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
yrStrt = 1979
yrLast = 1996
ymStrt = yrStrt*100 + 1
ymLast = yrLast*100 + 12
ff=addfile("/disk3/erainterim2.5/monthly/pressure/air.mon.mean.nc","r")
TIME = ff->time
YYYY = cd_calendar(TIME,-1)/100 ; entire file
iYYYY = ind(YYYY.ge.yrStrt .and. YYYY.le.yrLast)
ST=ff->t(iYYYY,{850},:,:)
STT=short2flt(ST)-273.15
copy_VarMeta(ST, STT)
STT@units= "C"
STT2=month_to_season(STT,"DJF")
f=addfile("/disk3/erainterim2.5/monthly/pressure/hgt.mon.mean.nc","r")
Z=f->z(0:215,:,:,:)
Z:=short2flt(Z)
ZDJF=month_to_season(Z, "DJF")
TT=ZDJF(:,{200},:,:)-ZDJF(:,{850},:,:)
copy_VarMeta(ZDJF(:,0,:,:), TT)
printVarSummary(TT)
printMinMax(TT, False)
neof = 1 ; number of EOFs
optEOF = True
optEOF@jopt = 1 ; This is the default; most commonly used; no need to specify.
;optEOF@jopt = 1 ; **only** if the correlation EOF is desired
optETS = False
; =================================================================
TT2 = TT(:,{-20:20},:)
rad = get_r2d("float")
clat = TT2&latitude
clat = sqrt( cos(rad*clat) ) ; gw for gaussian grid
printVarSummary(clat)
; =================================================================
; weight all observations
; =================================================================
xw = TT2*conform(TT2, clat, 1)
copy_VarMeta(TT2, xw)
xw@long_name = "hgt: "+Z@long_name
; =================================================================
eof = eofunc_n_Wrap(xw, neof, optEOF, 0)
eof=-1*eof ; *special* match sign of CPC
eof2 =eofunc_n_Wrap(STT2,neof,optEOF,0)
eof2=-1*eof2
eof_ts = eofunc_ts_n_Wrap (xw, eof, optETS, 0)
printVarSummary( eof2 ) ; examine EOF variables
eof_ts = dim_standardize_n( eof_ts, 0, 1) ; normalize
printMinMax(eof_ts,False)
; =================================================================
; Regress
; =================================================================
TT3=STT2(latitude|:,longitude|:,time|:)
tval=new((/73,144/),"float")
nxy =new((/73,144/),"integer")
lon = lonGlobeF(144, "lon", "lon", "degrees_east")
tval!1= "lon"
tval&lon = lon
lat = latGlobeF(73, "lat", "lat", "degrees_north")
tval!0= "lat"
tval&lat = lat
lon = lonGlobeF(144, "lon", "lon", "degrees_east")
nxy!1= "lon"
nxy&lon = lon
lat = latGlobeF(73, "lat", "lat", "degrees_north")
nxy!0= "lat"
nxy&lat = lat
eof_regres = eof2 ; create an array w meta data
do ne=0,neof-1
eof_regres(ne,:,:) = regcoef(eof_ts(ne,:),TT3, tval,nxy )
end do
; df=16
; tval=student_t(tval, df)
; printVarSummary(eof_regres)
;printMinMax(eof_regres,False)
printMinMax(tval, False)
;============================================================
; PLOTS
;============================================================
wks = gsn_open_wks("pdf","/disk1/dzz/study/ncl/TT/SATEOF") ; send graphics to PNG file
gsn_define_colormap(wks,"amwg256")
plot = new(neof,graphic) ; create graphic array
plot2= new(neof,graphic) ; only needed if paneling
; EOF patterns
res = True
res@gsnDraw = False ; don't draw yet
res@gsnFrame = False ; don't advance frame yet
;res@gsnAddCyclic = False ; plotted dataa are not cyclic
res@mpFillOn = False ; turn off map fill
res@mpOutlineOn = True
res@mpMinLatF = -40 ; zoom in on map
res@mpMaxLatF = 80
res@mpMinLonF = 0
res@mpMaxLonF = 360
res@mpCenterLonF = 180
res@cnFillOn = False ; turn on color fill
res@cnLinesOn = True ; True is default
res@cnLineLabelsOn = False ; True is default
res@gsnContourNegLineDashPattern = 1
res@cnLineDrawOrder ="PostDraw"
;res@cnFillPalette = "BlWhRe" ; set color map
res@cnLevelSelectionMode = "ManualLevels"
res@cnMinLevelValF = -1.
res@cnMaxLevelValF =1.5
res@cnLevelSpacingF = 0.5
res@lbLabelBarOn = False ; turn off individual lb's
res@cnInfoLabelOn = False
;res@lbTitleOn = False
res@cnSmoothingOn = True
;res@cnRasterSmoothingOn = True
;symMinMaxPlt(eof_regres, 0.5, False, res) ; contributed.ncl
;res2@cnRasterSmoothingOn = True
; panel plot only resources
resP = True ; modify the panel plot
resP@gsnMaximize = True ; large format
resP@gsnPanelLabelBar = True ; add common colorbar
resP@txString = ""
res2 = True
res2@gsnDraw = False
res2@gsnFrame = False
res2@cnFillOn = True ; turn on color fill
res2@cnLinesOn = False ; True is default
res2@cnLineLabelsOn = False ; True is default
;res2@gsnContourNegLineDashPattern = 1
;res@cnFillPalette = "BlWhRe" ; set color map
res2@lbLabelBarOn = False ; turn off individual lb's
res2@cnLevelSelectionMode = "ExplicitLevels"
res2@cnLevels = (/-2.120,-1.746,1.746,2.120/)
res2@cnFillColors = (/22,7,0,177,225/)
res2@cnInfoLabelOn = False
;res2@lbTitleOn = False
;res2@cnSmoothingOn = True
;*******************************************
; first plot
;*******************************************
do n=0,neof-1
res@gsnLeftString = "(a) T850"
res@gsnRightString = "CI=0.5 C"
plot(n)= gsn_csm_contour_map(wks,eof_regres(n,:,:),res)
plot2(n)=gsn_csm_contour(wks,tval,res2)
overlay(plot, plot2)
draw(plot)
frame(wks)
end do
gsn_panel(wks,plot,(/neof,1/),resP) ; now draw as one plot)
end

复制代码

我是小猪 · 发表于 2017-4-5 10:45:20

自顶！我自己找到原因啦，原来是等值线图的eof_regres的经纬度信息是（latitude，longitude），而填色图显著性检验的tval的经纬度信息是（lat,lon),然后我把tval的经纬度信息也改成（latitude，longitude），画出来的就是在最值区域里通过显著性检验了，不过我觉得tval的经纬信息（-90，90）/（0，357.5）和eof_regres的是一样的，可为什么NCL在识别的时候要按lat/latitude这样来呢，搞不懂

lgz · 发表于 2017-4-7 15:10:04

识别的时候要按lat/latitude

南国的佳人 · 发表于 2017-4-8 09:48:13

{:eb315:}{:eb315:}

我是小猪 · 发表于 2017-4-8 15:21:48

lgz 发表于 2017-4-7 15:10
识别的时候要按lat/latitude

原来是这样，谢啦！

Dannyqiu · 发表于 2018-7-17 11:25:38

可以请教一下楼主和大神们，程序里面的这行
TT=ZDJF(:,{200},:,:)-ZDJF(:,{850},:,:)
计算的是什么吗？

		自动登录	找回密码
密码			立即注册

[作图] 为什么变化最大的正负闭合中心通不过显著性检验

登录后查看更多精彩内容~