Link Search Menu Expand Document
Focus on Air Quality

整併8個批次wrfout成為全月檔案

Table of contents

背景

  • 全月的空品模擬會減省許多不必要的初始化時間，減少檔案個數，一如CAMx全月模擬的架構。
  • 然而wrfout中有不少變數、標籤是以批次起始時間為準的。wrfout無法靠簡單的ncrcat予以整併。
  • 變數名稱與內容如下表
變數名稱變數內容
ACGRDFLXACCUMULATED GROUND HEAT FLUX
ACSNOMACCUMULATED MELTED SNOW
RAINCACCUMULATED TOTAL CUMULUS PRECIPITATION
RAINSHACCUMULATED SHALLOW CUMULUS PRECIPITATION
RAINNCACCUMULATED TOTAL GRID SCALE PRECIPITATION
SNOWNCACCUMULATED TOTAL GRID SCALE SNOW AND ICE
GRAUPELNCACCUMULATED TOTAL GRID SCALE GRAUPEL
HAILNCACCUMULATED TOTAL GRID SCALE HAIL
ACHFXACCUMULATED UPWARD HEAT FLUX AT THE SURFACE
ACLHFACCUMULATED UPWARD LATENT HEAT FLUX AT THE SURFACE
  • 標籤如下表
標籤名稱標籤內容
XTIMEminutes since nc.SIMULATION_START_DATE
ITIMESTEPtimesteps since nc.SIMULATION_START_DATE
  • 全域屬性
屬性名稱屬性內容範例
SIMULATION_START_DATE批次開始時間2018-03-31_00:00:00
JULYR批次開始年代2018
JULDAY批次開始日期90

add_xtime程式說明

分段說明

  • 調用模組
     1	import netCDF4
     2	import subprocess
     3	import datetime
     4	import numpy as np
     5	import sys,os
     6	from calendar import monthrange
  • 由所作目錄讀取年月訊息、標籤、變數等等基本定義
     7	#working under 16??, the file in directory(or file by link) will be modified
     8	yrmn=subprocess.check_output('pwd',shell=True).decode('utf8').strip('\n').split('/')[-1]
     9	begd=datetime.datetime(2000+int(yrmn[:2]),int(yrmn[2:4]),1)+datetime.timedelta(days=-1)
    10	b,e=monthrange(2000+int(yrmn[:2]),int(yrmn[2:4]))
    11	x='XTIME'
    12	y='ITIMESTEP'
    13	#accumulation variables
    14	acc=['ACGRDFLX', 'ACSNOM', 'RAINC', 'RAINSH', 'RAINNC', 'SNOWNC', 'GRAUPELNC', 'HAILNC', 'ACHFX', 'ACLHF']
  • 由於程式將會更動acc變數的內容，建議先就wrfout的內容另存備份。
    15	#note acc should be saved and restored(if needed) before following actions:
    16	# for dm in 1 2 4;do
    17	#   for i in $(ls wrfout_d0${dm}*);do d=$(echo $i|cut -d'_' -f3)
    18	#     ncks -O -v Times,ACGRDFLX,ACSNOM,RAINC,RAINSH,RAINNC,SNOWNC,GRAUPELNC,HAILNC,ACHFX,ACLHF $i $d.nc;done
    19	#   ncrcat -O 2016*.nc acc_d0${dm}.nc
    20	# done
  • 每層網格系統的迴圈
    • 讀取起始時間、年、Julian Day數據
    • 累積變數由0起算
    21	for DM in ['1', '2','4']:
    22	  #each run must begin with same day(last day of previous month)
    23	  fname='wrfout_d0'+DM+'_'+begd.strftime("%Y-%m-%d")+'_00:00:00'
    24	  nc = netCDF4.Dataset(fname,'r')
    25	  min0=nc.variables[x][-1]+60
    26	  START_DATE=nc.SIMULATION_START_DATE
    27	  JULYR=nc.JULYR
    28	  JULDAY=nc.JULDAY
    29	  if JULYR%4==0:
    30	    JULDAY=min(366,JULDAY)
    31	  else:
    32	    JULDAY=min(365,JULDAY)
    33	  TITLE =nc.TITLE
    34	  # begin with zero accumulation
    35	  acmx={ac:np.zeros(shape=nc.variables[ac].shape) for ac in acc}
    36	  nc.close()
  • 批次的迴圈
    • 計算批次的起訖日期
    • 開啟檔案
    • 累積變數、紀錄以供下一批次起算。
    • 更改起算時間、年、Julian day
    37	  #run5~12
    38	  for r in range(5,13):
    39	    d0=(r-5)*4+2
    40	    dEnd=d0+4
    41	    if r==12:dEnd=d0+5    
    42	    for d in range(d0,dEnd):
    43	      nowd=(begd+datetime.timedelta(days=d)).strftime("%Y-%m-%d")
    44	      fname='wrfout_d0'+DM+'_'+nowd+'_00:00:00'
    45	      print(r,d,fname)
    46	      nc = netCDF4.Dataset(fname,'r+')
    47	      for ac in acc:
    48	        nc.variables[ac][:]+=acmx[ac]
    49	      if d==d0+3:
    50	        acmx={ac:nc.variables[ac][:] for ac in acc}
    51	      nc.SIMULATION_START_DATE=START_DATE
    52	      nc.START_DATE           =START_DATE
    53	      nc.JULYR                =JULYR
    54	      nc.JULDAY               =JULDAY
    55	      nc.TITLE                =TITLE
  • 逐時更改標籤數字
    56	      for t in range(24):
    57	        mins=min0+((d-1)*24+t)*60
    58	        nc.variables[x][t]=float(mins)
    59	        nc.variables[x].units='minutes since '+START_DATE
    60	        nc.variables[x].description='minutes since '+START_DATE
    61	        nc.variables[y][t]=int(mins)
    62	      nc.close()
    63

程式下載