About PDF of forecast and observation

From: Waqar Younas <vickyqau_at_nyahnyahspammersnyahnyah>
Date: Sun May 26 2013 - 01:26:26 MDT

Hi everyone,
I have one question regarding calculating probability density functions for
observation and forecast. I chose Pacific North American index both for
observation and forecast. The data dimensions of my observation and
forecast are (120,88) and (120,88,27) where 120 are lead time, 88 are
number of predictions and 27 are ensemble members. I used this code to
calculate the pdfs

;***********************************************************

; pdf_1.ncl

;

; Concepts illustrated:

; - Generating univariate probability distributions

; - Generating PDFs of each sample distribution

; - Paneling two plots horizontally on a page

;***********************************************************

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

  a = obs ;;;; dimensions are 120,88

  b = forecast ;;;;;;; dimensions are 120,88,27

  ap = pdfx(a, 0, False) ; default number of bins

  bp = pdfx(b, 0, False)

;------------------------------------------------

; Plot each PDF

;------------------------------------------------

  nVar = 2

  plot = new ( nVar, "graphic")

  wks = gsn_open_wks ("ps","pdf")

  res = True

  res@gsnDraw = False

  res@gsnFrame = False

  res@xyLineThicknessF = 2

  res@tiYAxisString = "PDF (%)"

  plot(0) = gsn_csm_xy (wks, ap@bin_center, ap, res)

  plot(1) = gsn_csm_xy (wks, bp@bin_center, bp, res)

  resP = True

  resP@txString = "obs and forecast pdfs"

  resP@gsnPanelRowSpec = True ; tell panel what order to
plt

  gsn_panel(wks,plot,(/1,2/),resP)

end

My question is how the function pdfx works because in the pdf1.ncl it is
not mentioned what dimension of x should and if x has more than 1 dimension
how it works (i.e. which dimension or it combines all dimensions)?
For forecast, the pdf should be at least calculated atleast for each lead
time by combining the predictions and ensemble members.

Thank you in advance.

-- 
Best Regards
Waqar Younas
PhD Candidate
Natural Resources and Environmental Studies (NRES)
University of Northern British Columbia (UNBC)
Canada

_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Sun May 26 01:26:36 2013

This archive was generated by hypermail 2.1.8 : Tue Jun 11 2013 - 12:03:58 MDT