def load_nclatlon(field, last='last100',includeyr1=False,verb=False,local=False):
""" return lat,lon from given field nc file. Uses preipreiice casename
if local=True, just gets lat/lon from DJF mean files (full globe...). ignores 'field', 'last'
"""
timepers,timesels=get_timeinfo(last,includeyr1)
casename = 'preipreiice'
if local:
#bd='/Users/kelly/DropboxWork/UWashCanSISE/DATA/relaxationruns/' #limited fields here
bd='/Users/kelly/DATA/DataDisk/' # limited but not shared so there are more data
fname=bd+'preipreiice/preipreiice_st_2922-3121_DJF_mean.nc'
else:
fname= basedir + casename +'/ts/' + casename + '_' + field + '_' + timepers[casename] + '_ts.nc'
if verb:
print fname
tmplat=cnc.getNCvar(fname,'lat')
if verb:
print 'lat.shape ' + str(tmplat.shape)
tmplon=cnc.getNCvar(fname,'lon')
if verb:
print 'lon.shape ' + str(tmplon.shape)
return tmplat,tmplon
def load_nclatlon(field, last='last100',includeyr1=False,verb=False):
""" return lat,lon from given field nc file. Uses preipreiice casename
"""
timepers,timesels=get_timeinfo(last,includeyr1)
casename = 'preipreiice'
fname= basedir + casename +'/ts/' + casename + '_' + field + '_' + timepers[casename] + '_ts.nc'
if verb:
print fname
tmplat=cnc.getNCvar(fname,'lat')
if verb:
print 'lat.shape ' + str(tmplat.shape)
tmplon=cnc.getNCvar(fname,'lon')
if verb:
print 'lon.shape ' + str(tmplon.shape)
return tmplat,tmplon
def load_nclev(field,verb=False):
timepers,timesels=get_timeinfo()
casename = 'preipreiice'
fname= basedir + casename +'/ts/' + casename + '_' + field + '_' + timepers[casename] + '_ts.nc'
if verb:
print fname
lev=cnc.getNCvar(fname,'plev')
return lev
def get_t63lat():
#plat = platform.system()
basepath = get_constantsbasepath()
#if plat == "Linux":
# basepath = '/HOME/rkm/work/DATA/CanAM4/constants/'
#else:
# #basepath = '/Users/kelly/CCCma/CanSISE/DATA/constants/' #@@
# basepath = '/Volumes/KellyDataDisk/work/DATA/CanESM2/' #@@@@ don't have the other dir on imac
fname = basepath + 't63_landmask.nc'
return cnc.getNCvar(fname,'lat')
def get_t63lev(): # prob isn't tied to t63..
#plat = platform.system()
basepath = get_constantsbasepath()
#if plat == "Linux":
# basepath = '/HOME/rkm/work/DATA/CanAM4/constants/'
#else:
# #basepath = '/Users/kelly/CCCma/CanSISE/DATA/constants/'
# basepath = '/Volumes/KellyDataDisk/work/DATA/CanESM2/' #@@@@ don't have the other dir on imac
fname = basepath + 'kem1rcp85a_v_001-061_climo.nc'
return cnc.getNCvar(fname,'plev')
def get_t63latbounds():
""" returns the lat_bnds from grid area file. Dims are [64,2]
"""
#plat = platform.system()
basepath = get_constantsbasepath()
#if plat == "Linux":
# basepath = '/HOME/rkm/work/DATA/CanAM4/constants/'
#else:
# #basepath = '/Users/kelly/CCCma/CanSISE/DATA/constants/' #@@
# basepath = '/Volumes/KellyDataDisk/work/DATA/CanESM2/' #@@@@ don't have the other dir on imac
fname = basepath + 'areacella_fx_CanESM2_historical-r1_r0i0p0.nc'
return cnc.getNCvar(fname,'lat_bnds')
def get_t63lon(remcyclic=False):
#plat = platform.system()
basepath = get_constantsbasepath()
#if plat == "Linux":
# basepath = '/HOME/rkm/work/DATA/CanAM4/constants/'
#else:
# #basepath = '/Users/kelly/CCCma/CanSISE/DATA/constants/' #@@
# basepath = '/Volumes/KellyDataDisk/work/DATA/CanESM2/' #@@@@ don't have the other dir on imac
fname = basepath + 't63_landmask.nc'
lon=cnc.getNCvar(fname,'lon')
if remcyclic:
# if an extra lon exists, remove it.
if np.mod(lon.shape[0],2) != 0: # if lon is odd, remove extra
lon=lon[:-1]
return lon
basepath = '/home/rkm/work/DATA/'
basepath2 = '/home/rkm/work/BCs/'
# read in HadISST SST & Thickness (and SIC for comparison)
# read in NSIDC SIC
timeper='2002-2011'
timeperc='1979-1989'
# NSIDC first
dset = 'NSIDC'
nsicfnamep = basepath2 + dset + '/nsidc_bt_128x64_1978m11_2011m12_sicn_' + timeper + 'climo.nc'
nsicfnamec = basepath2 + dset + '/nsidc_bt_128x64_1978m11_2011m12_sicn_' + timeperc + 'climo.nc'
nsicp = cnc.getNCvar(nsicfnamep,'SICN')
nsicc = cnc.getNCvar(nsicfnamec,'SICN')
# HadISST
dset='HadISST'
sstfnamep = basepath2 + dset + '/hadisst1.1_bc_128_64_1870_2013m03_gt_' + timeper + 'climo.nc'
sstfnamec = basepath2 + dset + '/hadisst1.1_bc_128_64_1870_2013m03_gt_' + timeperc + 'climo.nc'
sitfnamep = basepath2 + dset + '/hadisst1.1_bc_128_64_1870_2013m03_sic_' + timeper + 'climo.nc'
sitfnamec = basepath2 + dset + '/hadisst1.1_bc_128_64_1870_2013m03_sic_' + timeperc + 'climo.nc'
hsicfnamep = basepath2 + dset + '/hadisst1.1_bc_128_64_1870_2013m03_sicn_' + timeper + 'climo.nc'
hsicfnamec = basepath2 + dset + '/hadisst1.1_bc_128_64_1870_2013m03_sicn_' + timeperc + 'climo.nc'
hsicp = cnc.getNCvar(hsicfnamep,'SICN')
hsicc = cnc.getNCvar(hsicfnamec,'SICN')
# cmincg = 238; cmaxcg = 308 cmincg = -35 cmaxcg = 35 cmincnhg = -45 cmaxcnhg = 45 # polar NH cming = -2 cmaxg = 2 cminmg = -3 cmaxmg = 3 cmapg = "blue2red_20" fnamecs = basepath + casename + subdir + casename + "_" + fields + "_" + timstr + "_ts.nc" fnamep2s = basepath + casenamep2 + subdir + casenamep2 + "_" + fields + "_" + timstr + "_ts.nc" fldcs = cnc.getNCvar(fnamecs, fields.upper(), timesel="0002-01-01,0111-12-31") * convs fldp2s = cnc.getNCvar(fnamep2s, fields.upper(), timesel="0002-01-01,0111-12-31") * convs fnamecg = basepath + casename + subdir + casename + "_" + fieldg + "_" + timstr + "_ts.nc" fnamep2g = basepath + casenamep2 + subdir + casenamep2 + "_" + fieldg + "_" + timstr + "_ts.nc" fldcg = cnc.getNCvar(fnamecg, fieldg.upper(), timesel="0002-01-01,0111-12-31") * convg fldp2g = cnc.getNCvar(fnamep2g, fieldg.upper(), timesel="0002-01-01,0111-12-31") * convg lat = cnc.getNCvar(fnamecs, "lat") lon = cnc.getNCvar(fnamecs, "lon") fldcsclim, fldcsstd = cutl.climatologize3d(fldcs) fldpsclim, fldpsstd = cutl.climatologize3d(fldp2s) fldcgclim, fldcgstd = cutl.climatologize3d(fldcg)
cminn = -0.25
cmaxn = 0.25 # for nudge vs original
ylims = (0, 0.15)
conv = 1 / 913.0
fontP = fm.FontProperties()
fontP.set_size("small")
seasons = ("SON", "DJF", "MAM", "JJA")
flddt = {}
for sii, skey in enumerate(sims):
fname = basepath + skey + subdir + skey + "_" + field + "_" + timeperdt[skey] + "_ts.nc"
if sii == 0: # only get once
time = cnc.getNCvar(fname, "time", timesel=timeseldt[skey])
ntime = len(time)
lat = cnc.getNCvar(fname, "lat")
lon = cnc.getNCvar(fname, "lon")
flddt[skey] = cnc.getNCvar(fname, ncfield, timesel=timeseldt[skey]) * conv
# <headingcell level=3>
# SICN: Below are sea ice concentration comparisons for various combinations
# (figure titles are below figures)
# <codecell>
if showmaps:
fldc = flddt["iga"]
ncfield='PHI'
conv=1/con.get_g()
if field in ('gz','t','u','v','q'): # only 'gz' is implemented here officially @@
fieldstr=field+str(level/100) # for figure names
field=field+str(level) # for filename
fdict = {'field':field, 'ncfield': ncfield,
'conv': conv, 'fieldstr': fieldstr,
'level': level}
# ########### get baseline data: monthly and seasonal ##################
# get baseline to which to compare
fnamecc = basepath + cmpcasenamec + subdir + cmpcasenamec + '_' + field + '_001-121_ts.nc'
fnamepc = basepath + cmpcasenamep + subdir + cmpcasenamep + '_' + field + '_001-121_ts.nc'
fldcc = cnc.getNCvar(fnamecc,ncfield,timesel=timesel)*conv
fldpc = cnc.getNCvar(fnamepc,ncfield,timesel=timesel)*conv
flddc = fldpc-fldcc
lat = cnc.getNCvar(fnamecc,'lat')
lon = cnc.getNCvar(fnamecc,'lon')
# seasonal baseline to which to compare
fldccsea = np.zeros((4,len(lat),len(lon)))
fldpcsea = np.zeros((4,len(lat),len(lon)))
flddcsea = np.zeros((4,len(lat),len(lon)))
for seaii,sea in enumerate(seasons):
fldccsea[seaii,...] = np.mean(cnc.getNCvar(fnamecc,ncfield,timesel=timesel,seas=sea)*conv,axis=0)
fldpcsea[seaii,...] = np.mean(cnc.getNCvar(fnamepc,ncfield,timesel=timesel,seas=sea)*conv,axis=0)
flddcsea[seaii,...] = fldpcsea[seaii,...]-fldccsea[seaii,...]
def get_lon(local=False):
fnames=build_filenames({'field':'Z3500','comp':'cam'},'B20TRC5CNBDRD-BRCP85C5CNBDRD',verb=False,local=local)
return cnc.getNCvar(fnames[1],'lon')
## ccasename='historicalrcp45'
# loop through ens members:
cfldcall = np.zeros((ensnum,12)) # a climo
cfldpall = np.zeros((ensnum,163*12)) # hard-coded monthly timeseries 1850-2012
for eii in range(1,ensnum+1): # five ens members
cfnamec = basepath + ccasename + '/' + cfield + '/' + cfield +\
'_' + comp + '_' + cmodel + '_' + ccasename +\
'_r' + str(eii) + 'i1p1_' + ctimstr + 'climo.nc' # to take anomaly from
cfnamep = basepath + ccasenamep + '/' + cfield + '/' + cfield +\
'_' + comp + '_' + cmodel + '_' + ccasenamep +\
'_r' + str(eii) + 'i1p1_' + ctimeper + '.nc' # sic, 128x64
cfldc = cnc.getNCvar(cfnamec,cfield)*cconv # ctl time period avg
cfldp = cnc.getNCvar(cfnamep,cfield)*cconv # full timeseries
lat = cnc.getNCvar(cfnamec,'lat')
lon = cnc.getNCvar(cfnamec,'lon')
## cfldc = np.dstack((cfldc,cfldc[...,0])) # add wraparound lon
## cfldp = np.dstack((cfldp,cfldp[...,0])) # add wraparound lon
# if sea ice frac: calc area and save it
if cfield == 'sic':
# calc sea ice area
# mult fraction by grid cell area & sum
areas = cutl.calc_cellareas(lat,lon)
careasp = np.tile(areas,(cfldp.shape[0],1,1)) # need one per time
careasc = np.tile(areas,(cfldc.shape[0],1,1)) # need one per month (climo)
def get_lon(local=False):
fnames=build_filenames({'field':'tas','comp':'Amon'},'historical',verb=False,local=local)
return cnc.getNCvar(fnames[0],'lon')
def load_ncfield(field, ncfield, zonal=True,conv=1,mo=0,season=None,last='last100',
includeyr1=False,verb=False,local=False,seacyc=False,pulse=False):
"""
If zonal=True, compute zonal mean.
zonal can also be a region, defined in regiondict in constants.py
If local, use local data (which is limited)
If seacycle, return the 12 month climo average, which only exists
1. locally (as of Mar 24 2017) and 2. is only avail for the last200
returns dictionaries: full field w/ time, zonal mean w/ time
"""
timepers,timesels = get_timeinfo(last,includeyr1)
if local:
#bd='/Users/kelly/DropboxWork/UWashCanSISE/DATA/relaxationruns/' #limited fields here
bd='/Users/kelly/DATA/DataDisk/' # limited but not shared so there are more data
subdir='/'
else:
bd=basedir
subdir='/ts/'
seasonalizedt = {'mo':mo, 'season': season}
ncflddt={}; ncfldtmdt={}; ncfldzmdt={}
for casename in casenames:
if pulse==False and ('pulse' in casename): continue
if seacyc:
if last!='last200':
raise Exception('load_ncfield(): if seacyc=True, last must = last200')
if local!=True:
raise Exception('load_ncfield(): if seacyc=True, local must be True (as of Mar 24 2017)')
fname = bd + casename+subdir+casename + '_' + field + '_' + timepers[casename] + '_seacycle.nc'
else:
fname= bd + casename +subdir + casename + '_' + field + '_' + timepers[casename] + '_ts.nc'
timesel=timesels[casename]
if verb:
print fname
try:
fld=cnc.getNCvar(fname,ncfield,timesel=timesel)*conv # remlon not operational??? why commented out??
if verb:
print fname + ', fld.shape ' + str(fld.shape) # @@@
except IOError as rte: # previously this was a RuntimeError. Not sure what changed. 10/14/2016
print rte
if 'No such file or directory' in rte.args:
# Here the problem is most likely that pico2ipulse filename fails b/c
# there is no NH version of it (for when last200=True). Remove the NH and try again.
# @@ Of course this could be a problem too b/c now we have full globe for just this run.
if field[-2:] != 'NH':
print 'not an NH versus global problem...File really not found'
raise
fname= basedir + casename +'/ts/' + casename + '_' + field[:-2] + '_' + timepers[casename] + '_ts.nc'
print 'Try again: ' + fname
fld=cnc.getNCvar(fname,ncfield,timesel=timesel)*conv
tmplat=cnc.getNCvar(fname,'lat')
if last=='last200': # how to tell which spot lat is?
if zonal: # then it is likely time x lat x lon
#print 'zonal is true. fld dims? ' + str(fld.shape)
fld=fld[...,tmplat>0,:]
else: # likely time x lev x lat
fld=fld[...,tmplat>0]
else:
raise
# just get the full timeseries, and zonal mean if asked.
if 1: #timeseries: # do not time average (actually save time average anyway)
if (season==None and mo==0):
pass
else:
fld = cutl.seasonalize(fld,**seasonalizedt)
#fldtm = fld
# zonal average
if zonal==True:
fldzm = fld[...,:-1].mean(axis=2) # remove extra lon
elif zonal in con.get_regiondict().keys():
print 'lat,lon shapes ' + str(lat.shape),str(lon.shape)
tempo, _ = cutl.mask_region(fld,lat,lon,region=zonal)
fldzm = tempo.mean(axis=2)
else:
fldzm = None
#fldtm = fldtmzm.mean(axis=0)
# else:
# # seasonal & time mean
# fldtm = np.mean(cutl.seasonalize(fld,**seasonalizedt),axis=0)
# # zonal average
# if zonal:
# fldtmzm = fldtm[...,:-1].mean(axis=1)
# elif zonal in con.get_regiondict().keys():
# tempo, _ = cutl.mask_region(fldtm,lat,lon,region=zonal)
# fldtmzm = tempo.mean(axis=1)
# else:
# fldtmzm = fldtm
print 'fld.shape: ' + str(fld.shape)
if fldzm!=None:
print 'fldzm.shape: ' + str(fldzm.shape)
ncflddt[casename] = fld # keep all dims
#ncfldtmdt[casename] = fldtm # time mean
ncfldzmdt[casename] = fldzm # zonal mean (w/ time dim)
# try this: add coords # @@@@@@@@@@
lat,lon = load_nclatlon(field,last=last,includeyr1=includeyr1,verb=verb,local=local)
ncflddt['lat'] = lat; ncflddt['lon'] = lon;
try:
lev = load_nclev(field,last=last,includeyr1=includeyr1,verb=verb)
ncflddt['lev'] = lev
except:
print "No lev coord. Leave it"
ncfldzmdt['lat'] = lat
return ncflddt, ncfldzmdt
plat = platform.system()
if plat == 'Darwin': # means I'm on my mac
basepath2 = '/Volumes/MyPassport1TB/DATA/CanSISE/'
basepath = '/Users/kelly/CCCma/CanSISE/BoundaryConditionFiles/' # @@ not sure
else: # on linux workstation in Vic
basepath = '/home/rkm/work/BCs/'
if checkpert1s:
# New Pert1 dataset that checks for open water colder than 271.2
pert1newf = 'GTfrzchkhistoricalrcp452002-2012_BC_CanESM2_historical_1979-1989_1870010100-2010120100.nc'
lat = cnc.getNCvar(pert1newf,'lat')
lon = cnc.getNCvar(pert1newf,'lon')
pert1n = cnc.getNCvar(pert1newf,'GT')
pert1n = pert1n[0:12,...]
# original BC files for PERT1. I expected that the ssts would be below freezing in some places
# where SICN was gone in PERT1 but not control.
pert1f = '/home/rkm/work/BCs/CanESM2/GT_BC_CanESM2_historical1979-1989_1870010100-2011020100.nc'
pert1=cnc.getNCvar(pert1f,'GT')
pert1 = pert1[0:12,...]
pert1sicf = '/home/rkm/work/BCs/CanESM2/SICN_BC_CanESM2_historical2002-2012_1870010100-2011020100.nc'
def loaddata(fields, simulations, ncfields=None,model='CanAM4',timeper='001-121',timefreq=None,
levsel=None,meantype=None,filetype='diff',region=None,alsomask=None,rettype='dict'):
""" loaddata(fields, simulations,ncfields=None,model='CanAM4',timeper='001-121',timefreq=None,
levsel=None, meantype=None,filetype='diff',region=None)
fields: tuple of field names [@@update. only one field now but still need tuple]
simulations: tuple of simulation names (diff names='E1'...'ENS','NSIDC' etc)
ncfields: tuple of ncfield names (var name in file itself). Default to upper case of field
model: for now only 'CanAM4' is implemented
timeper: time period of the data (used for filename). default '001-121'
timefreq: time frequency TO RETURN. default all data
'monthly'|'seasonal'|'climo'|'ANN'|'DJF'|'JJA'|'NDJ'|'MAM'|'SON'|
1,2,3,4,5,6,7,8,9,10,11,12
levsel: select level in Pa (e.g. 50000 for 500hPa). default all levels
meantype: 'time','zonal' @@for now. default None, but recommended to choose
one if loading multiple variables and multiple simulations at once.
It is assumed that time is the first dimension.
filetype: 'diff','ctl','pert','pval'
Default is diff where both ctl and pert are read in and differenced.
region: any of the regions in constants -> region dict. default None.
alsomask: if specified as 'land' or 'ocean', then calc_regmean() will mask before
computing regional avg (ie. will NOT include it in average).
Only used if region!=None. Default None.
returns: nested dictionary@@
FIELDS->SIMULATIONS->TIMEFREQ
Load requested fields from requested CanAM4 simulations
into dictionaries (dataframes?). Function automatically
skips first year of simulation and gets a timeseries (or climo
if requested) of the rest (assumed through year 121).
3D data and 'turb' not yet implemented! @@
"""
if model!='CanAM4':
print 'model not supported!'
return -1
print '@@ probably should invert the order such that it is field, season, sim?'
#bp=con.get_basepath()
#basepath=bp['basepath'] + model + '/'; subdir=bp['subdir']
timesel='0002-01-01,0121-12-31'
seabool=False # set to True if requested time freq is one season or climo
monbool=False # set to True if requested time freq is one month
if timefreq=='monthly':
# get all months
tf = con.get_mon()
elif timefreq=='seasonal':
# get all 4 seasons
tf = 'DJF','MAM','JJA','SON'
elif timefreq in range(1,13):
# choose an individual month
tf=timefreq
monbool=True
elif timefreq in ('climo','ANN','DJF','JJA','NDJ','MAM','SON','ND','JF','SO'):
tf=timefreq
seabool=True
print tf # @@
# @@@@@ add handling of sia!
#datadict = dict.fromkeys(fields,{})
#for fii,field in enumerate(fields):
if 1: # GET RID OF FIELD dim too 5/8/2015
fii=0; field=fields[0]
if ncfields==None:
ncfield=field.upper()
else:
ncfield=ncfields[fii]
print field,ncfield #@@
# assume simulations entered are of the form: E1, R3, ENSE etc. Then
# filetype input arg tells which simulation to get (or both)
simdict = dict.fromkeys(simulations,{})
for sim in simulations:
print sim # @@
#timdict = dict.fromkeys(tf)
# construct filename here @@
fnamec,fnamep = con.build_filepathpair(sim,field)
#fname = basepath+sim+subdir+sim+'_'+field+'_'+timeper+'_ts.nc'
print fnamec
tfkey = tf
#for tfkey in tf:
# print tfkey # @@
#print 'too many levels in the dict...get rid of seasonal keys and just do one@@@ 5/1/2015'
# @@ get the data with cnc.getNCvar() here
ncparams = {}
if monbool:
ncparams = {'monsel': tfkey}
elif seabool:
ncparams = {'seas': tfkey}
if levsel!=None:
ncparams['levsel'] = levsel
if meantype=='zonal':
ncparams['calc'] = 'zm'
if filetype=='diff' or filetype=='pval':
if field in ('turb','net'):
fnamec,fnamep = con.build_filepathpair(sim,'hfl')
fnamecb,fnamepb = con.build_filepathpair(sim,'hfs')
ctl = cnc.getNCvar(fnamec,'HFL',timesel=timesel,**ncparams) +\
cnc.getNCvar(fnamecb,'HFS',timesel=timesel, **ncparams)
pert = cnc.getNCvar(fnamep,'HFL',timesel=timesel,**ncparams) +\
cnc.getNCvar(fnamepb,'HFS',timesel=timesel,**ncparams)
if field=='net':
fnamecc,fnamepc = con.build_filepathpair(sim,'flg')
ctl = ctl - cnc.getNCvar(fnamecc,'FLG',timesel=timesel,**ncparams)
pert = pert - cnc.getNCvar(fnamepc,'FLG',timesel=timesel,**ncparams)
else:
pert = cnc.getNCvar(fnamep,ncfield,timesel=timesel,**ncparams)
ctl = cnc.getNCvar(fnamec,ncfield,timesel=timesel,**ncparams)
fld = pert - ctl
elif filetype=='ctl':
if field in ('turb','net'):
fnamec,fnamep = con.build_filepathpair(sim,'hfl')
fnamecb,fnamepb = con.build_filepathpair(sim,'hfs')
fld = cnc.getNCvar(fnamec,'HFL',timesel=timesel,**ncparams) +\
cnc.getNCvar(fnamecb,'HFS',timesel=timesel, **ncparams)
if field=='net':
fnamecc,fnamepc = con.build_filepathpair(sim,'flg')
fld = fld - cnc.getNCvar(fnamecc,'FLG',timesel=timesel,**ncparams)
else:
fld = cnc.getNCvar(fnamec,ncfield,timesel=timesel,**ncparams)
elif filetype=='pert':
if field in ('turb','net'):
fnamec,fnamep = con.build_filepathpair(sim,'hfl')
fnamecb,fnamepb = con.build_filepathpair(sim,'hfs')
fld = cnc.getNCvar(fnamep,'HFL',timesel=timesel,**ncparams) +\
cnc.getNCvar(fnamepb,'HFS',timesel=timesel,**ncparams)
if field=='net':
fnamecc,fnamepc = con.build_filepathpair(sim,'flg')
fld = fld - cnc.getNCvar(fnamepc,'FLG',timesel=timesel,**ncparams)
else:
fld = cnc.getNCvar(fnamep,ncfield,timesel=timesel,**ncparams)
else:
print "filetype not supported! ['diff'|'ctl'|'pert'|'pval']"
return -1
if region != None:
lat=cnc.getNCvar(fnamec,'lat'); lon=cnc.getNCvar(fnamec,'lon')
if filetype=='pval':
pert = cutl.calc_regmean(pert,lat,lon,region,alsomask=alsomask)
ctl = cutl.calc_regmean(ctl,lat,lon,region,alsomask=alsomask)
(tstat,pval) = cutl.ttest_ind(pert,ctl)
fld=pval
else:
fld = cutl.calc_regmean(fld,lat,lon,region,alsomask=alsomask)
if meantype=='time':
if filetype=='pval':
# this is an error. filetype supercedes meantype so time avg won't be done
print 'filetype=pval and meantype-time. Ignore meantype and return pvals @@'
timdict=fld
else:
#fldstd = np.std(fld,axis=0)
fld = np.mean(fld,axis=0)
#timdict[tfkey] = fld #,fldstd
timdict=fld
else:
#timdict[tfkey] = fld
timdict=fld
simdict[sim] = timdict
#datadict[field]=simdict
# can I set attributes to a dictionary? @@ like for nfields, nsims, ntimes?
#return datadict
if rettype=='ndarray':
# convert the dict to an array:
# get the last sim data to initialize ndarray
initshape=simdict[sim].shape
initshape=(len(simulations),) + initshape
tmp=np.zeros(initshape)
for sii,skey in enumerate(simulations):
tmp[sii,:] = simdict[skey]
return tmp
else:
return simdict
def load_LEdata(fielddict, ens, seas=None, timesel=None,infodict=None,ftype='fullts',
calctype=None, calcdict=None, rettype='dict',conv=1, region=None,local=False,
subens=False,verb=True,zonal=False):
"""
ftype: type of filename to build. Default 'fullts' for full timeseries (assumes 192001-202012)
Also can override default time period of file by setting ftype to timeperiod.
e.g. ftype = '192001-200512'
seas has to be a tuple (but will only get one season)
rettype: 'dict' or 'ndarray' as return type.
default is dict of DataFrames (to make a Panel).
else, 'ndarray' is a 4D array
local: is data in ~/work/DATA (local=True) or /raid/rc40
subens: overrides getting all ensemble members. Must be tuple (e.g. (5,10) get ens mems 5 thru 9)
@@ add regional avg functionality?
returns an object of type rettype
"""
field = fielddict['field']
ncfield=fielddict['ncfield']
comp = fielddict['comp']
flist = build_filenames(fielddict, ens,ftype=ftype,timesel=timesel,local=local,
subens=subens,verb=verb)
fname1 = flist[0]
if verb:
print ' @@ fname1 ' + fname1
if rettype=='dict':
fldret={}
elif rettype=='ndarray':
if zonal:
tmp = cnc.getNCvar(fname1,ncfield,timesel=timesel,calc='zm')*conv
else:
tmp = cnc.getNCvar(fname1,ncfield,timesel=timesel)*conv
if len(tmp.shape)>3:
print '4D variable not supported! @@@@'
return -1
elif len(tmp.shape)==3:
# first dim is time, reshape the 2nd two into one spatial dim
n0=tmp.shape[0]
n1=tmp.shape[1]
n2=tmp.shape[2]
initshape=(len(flist), n0, n1*n2) # 3D, ens x time x space
rshape=initshape[1:] # what to reshape individual data to
elif len(tmp.shape)==2:
initshape=list(tmp.shape)
initshape.insert(0,len(flist))
initshape=tuple(initshape) # 3D, ens x time x space
rshape=None
elif len(tmp.shape)==1:
initshape = (len(flist),tmp.shape[0])
rshape=None
print initshape
fldret = np.zeros(initshape)
print fldret.shape # @@@
else:
raise Exception('return type not supported!! @@@')
return -1
numens = len(flist)
#fldall = np.zeros(numens)
#dlist=[]
if rettype=='dict':
timedim = cnc.getNCvar(fname1,'time',timesel=timesel)
if verb: print timedim
for eii,fname in enumerate(flist):
if seas==None:
# don't average, return all months
if zonal:
fld = cnc.getNCvar(fname,ncfield,timesel=timesel,calc='zm')*conv
else:
fld = cnc.getNCvar(fname,ncfield,timesel=timesel)*conv
#if region!=None:
# lat=cnc.getNCvar(fname,'lat')
# lon=cnc.getNCvar(fname,'lon')
# fld = cutl.calc_regmean(fld,lat,lon,region=region)
# do calcs:
if calctype!=None:
print 'calcs not implemented yet!! @@@'
if rettype=='ndarray':
fldseas=fld.reshape(rshape)
else:
if len(fld.shape)!=3:
raise Exception('load CESM LE with rettype dict will break if dim of var is not 3 @@@')
return -1
n0,n1,n2=fld.shape
fldseas=fld.reshape((n0,n1*n2)) # flattening
else:
fldseas={}
for sea in seas:
if zonal:
# @@@ not expecting this to work.
fld = cnc.getNCvar(fname,ncfield,timesel=timesel,seas=sea,calc='zm')*conv
else:
fld = cnc.getNCvar(fname,ncfield,timesel=timesel,seas=sea)*conv
# do calcs
if calctype!=None:
print 'calcs not implemented yet!! @@@'
fldseas[sea]=fld
if rettype=='dict':
fldret[fname]= pd.DataFrame(fldseas,index=timedim)
elif rettype=='ndarray':
if seas!=None:
print 'ndarray return type with a seasonal avg does not work yet @@@@'
fldret[eii,...] = fldseas
#dlist.append(fldseas)
return fldret
def load_ncfield(field, ncfield, zonal=True,conv=1,mo=0,season=None,last='last100',includeyr1=False,verb=False):
"""
If zonal=True, compute zonal mean.
zonal can also be a region, defined in regiondict in constants.py
returns dictionaries: full field w/ time, zonal mean w/ time
"""
timepers,timesels = get_timeinfo(last,includeyr1)
seasonalizedt = {'mo':mo, 'season': season}
ncflddt={}; ncfldtmdt={}; ncfldzmdt={}
for casename in casenames:
fname= basedir + casename +'/ts/' + casename + '_' + field + '_' + timepers[casename] + '_ts.nc'
timesel=timesels[casename]
if verb:
print fname
try:
fld=cnc.getNCvar(fname,ncfield,timesel=timesel)*conv # remlon not operational??? why commented out??
if verb:
print fname + ', fld.shape ' + str(fld.shape) # @@@
except IOError as rte: # previously this was a RuntimeError. Not sure what changed. 10/14/2016
print rte
if 'No such file or directory' in rte.args:
# Here the problem is most likely that pico2ipulse filename fails b/c
# there is no NH version of it (for when last200=True). Remove the NH and try again.
# @@ Of course this could be a problem too b/c now we have full globe for just this run.
if field[-2:] != 'NH':
print 'not an NH versus global problem...File really not found'
raise
fname= basedir + casename +'/ts/' + casename + '_' + field[:-2] + '_' + timepers[casename] + '_ts.nc'
print 'Try again: ' + fname
fld=cnc.getNCvar(fname,ncfield,timesel=timesel)*conv
tmplat=cnc.getNCvar(fname,'lat')
if last=='last200': # how to tell which spot lat is?
if zonal: # then it is likely time x lat x lon
#print 'zonal is true. fld dims? ' + str(fld.shape)
fld=fld[...,tmplat>0,:]
else: # likely time x lev x lat
fld=fld[...,tmplat>0]
else:
raise
# just get the full timeseries, and zonal mean if asked.
if 1: #timeseries: # do not time average (actually save time average anyway)
if (season==None and mo==0):
pass
else:
fld = cutl.seasonalize(fld,**seasonalizedt)
#fldtm = fld
# zonal average
if zonal==True:
fldzm = fld[...,:-1].mean(axis=2) # remove extra lon
elif zonal in con.get_regiondict().keys():
print 'lat,lon shapes ' + str(lat.shape),str(lon.shape)
tempo, _ = cutl.mask_region(fld,lat,lon,region=zonal)
fldzm = tempo.mean(axis=2)
else:
fldzm = None
#fldtm = fldtmzm.mean(axis=0)
# else:
# # seasonal & time mean
# fldtm = np.mean(cutl.seasonalize(fld,**seasonalizedt),axis=0)
# # zonal average
# if zonal:
# fldtmzm = fldtm[...,:-1].mean(axis=1)
# elif zonal in con.get_regiondict().keys():
# tempo, _ = cutl.mask_region(fldtm,lat,lon,region=zonal)
# fldtmzm = tempo.mean(axis=1)
# else:
# fldtmzm = fldtm
print 'fld.shape: ' + str(fld.shape)
if fldzm!=None:
print 'fldzm.shape: ' + str(fldzm.shape)
ncflddt[casename] = fld # keep all dims
#ncfldtmdt[casename] = fldtm # time mean
ncfldzmdt[casename] = fldzm # zonal mean (w/ time dim)
return ncflddt, ncfldzmdt
+ "_"
+ timeperc
+ "climo.nc"
)
landmask = con.get_t63landmask()
if doobs:
pass # do not remove wraparound lon
else:
landmask = landmask[..., :-1]
landmask = np.tile(landmask, (12, 1, 1))
print filename
timefldp = cnc.getNCvar(filename, "time")
lat = cnc.getNCvar(filename, "lat")
lon = cnc.getNCvar(filename, "lon")
if doobs:
fldp = cnc.getNCvar(filename, bcfield)
fldc = cnc.getNCvar(filenamec, bcfield)
else:
fldp = cnc.getNCvar(filename, field)
fldc = cnc.getNCvar(filenamec, field) # we should only need this to set SSTs in near-future BCs
print filenamec
if field == "ts" and adjustsst:
# assume adjustsst will always be true now
def pattcorr_withinensemble(ename,fdict,latlim=60,timesel='0002-01-01,0121-12-31'):
""" pattcorr_withinensemble(ename,field,latlim=60)
pattern corr each member of ensemble with each other one
return pctable, pctablesea (DataFrames)
"""
# @@ need diffdict
field=fdict['field']
ncfield=fdict['ncfield']
conv=fdict['conv']
seasons=('SON','DJF','MAM','JJA')
if ename=='ANT':
ename='histIC'
elif ename=='TOT':
ename='histBC'
enssims = con.build_ensemblesims(ename)
ensnum=len(enssims)
print 'ENSSIMS: ' # @@@
print enssims # @@
# generate weights for the pattern corr
lat = con.get_t63lat()
lon = con.get_t63lon()
areas = cutl.calc_cellareas(lat,lon)
areas = areas[lat>latlim,:]
weights = areas / np.sum(np.sum(areas,axis=1),axis=0)
# ========= create diffdict first =====
diffdict={}
seadiffdict={}
for skey in enssims:
fnamec,fnamep = con.build_filepathpair(skey,field)
# monthly calc
fldc = cnc.getNCvar(fnamec,ncfield,timesel=timesel)*conv
fldp = cnc.getNCvar(fnamep,ncfield,timesel=timesel)*conv
fldd = fldp-fldc
# Monthly
flddclimo,flddstd = cutl.climatologize(fldd) # climo first (don't need to do for BCs technically)
#flddcclimo,flddcstd = cutl.climatologize(flddc) # climo first. baseline diff data
diffdict[skey] = flddclimo
print skey + ' ' + str(flddclimo.shape) # @@@
# Seasonal
flddsea = np.zeros((4,len(lat),len(lon)))
for seaii,sea in enumerate(seasons):
fldcsea = np.mean(cnc.getNCvar(fnamec,ncfield,timesel=timesel,seas=sea)*conv,axis=0)
fldpsea = np.mean(cnc.getNCvar(fnamep,ncfield,timesel=timesel,seas=sea)*conv,axis=0)
flddsea[seaii,...] = fldpsea-fldcsea
seadiffdict[skey] = flddsea
# ======= Now do pattern corrs within ensemble ====
# ======= copied from Canam4_BCpatterncorr-Copy0.py ===========
outterdict= dict.fromkeys(enssims)
for skey1 in enssims:
outfld = diffdict[skey1]
innerdict = dict.fromkeys(enssims)
for skey2 in enssims:
#print skey1 + ' compared to ' + skey2
infld = diffdict[skey2]
# for each month, compute pattern corr
pc = np.zeros((12))
for mii,mon in enumerate(con.get_mon()):
tmp = np.squeeze(infld[mii,lat>latlim,...])
tmpcmp = np.squeeze(outfld[mii,lat>latlim,...])
pc[mii] = cutl.pattcorr(tmp.flatten()*weights.flatten(),
tmpcmp.flatten()*weights.flatten())
innerdict[skey2] = pc
outterdict[skey1] = innerdict
pctable = pd.DataFrame(outterdict) # 5x5
# seasonal
outterdictsea= dict.fromkeys(enssims)
for skey1 in enssims:
outfld = seadiffdict[skey1]
innerdictsea = dict.fromkeys(enssims)
for skey2 in enssims:
#print skey1 + ' compared to ' + skey2
infld = seadiffdict[skey2]
# for each season, compute pattern corr
pcsea = np.zeros((4))
for seaii,sea in enumerate(seasons):
tmp = np.squeeze(infld[seaii,lat>latlim,...])
tmpcmp = np.squeeze(outfld[seaii,lat>latlim,...])
pcsea[seaii] = cutl.pattcorr(tmp.flatten()*weights.flatten(),
tmpcmp.flatten()*weights.flatten())
innerdictsea[skey2] = pcsea
outterdictsea[skey1] = innerdictsea
pctablesea = pd.DataFrame(outterdictsea) # 5x5
return pctable, pctablesea
def load_LEdata(fielddict, ens, seas=None, timesel=None,infodict=None,ftype='fullts',
calctype=None, calcdict=None, rettype='dict',conv=1, region=None,local=True,
orig=None,subens=False,verb=True,zonal=False):
""" def loadLEdata(fielddict, seas=('DJF','MAM','JJA','SON'), timesel=None, infodict=None, calctype=None, calcdict=None)
ftype: type of filename to build. Default 'fullts' for full timeseries.
'fullclimo' for '1950-2020_climo' or, given timesel:
for styr-enyr_climo. or 'ensmean'.
Also can override default time period of file by setting ftype to timeperiod.
e.g. ftype = '195001-201012'
seas has to be a tuple
rettype: 'dict' or 'ndarray' as return type.
default is dict of DataFrames (to make a Panel).
else, 'ndarray' is a 4D array
local: is data in /Volumes/KellyDataDisk/home/work/DATA (local=True) or /raid/ra40
orig='just' means just load the original 5 (historical+rcp85) runs (consistent w/ LE).
orig='just45' means just load original 5 (historical+rcp45) to the full LE.
orig='add' means add the original 5 (historical+rcp85) to the full LE (consistent w/ LE)
orig='add45' means add the original 5 (historical+rcp45) to the full LE.
subens: overrides getting all sim groups in ensemble. must be tuple. e.g. ('historical-r1',)
@@ add regional avg functionality?
returns an object of type rettype
"""
field = fielddict['field']
ncfield=fielddict['ncfield']
comp = fielddict['comp']
flist = build_filenames(fielddict, ens,ftype=ftype,timesel=timesel,local=local,
orig=orig,subens=subens,verb=verb)
fname1 = flist[0]
if verb:
print ' @@ fname1 ' + fname1
timedim = cnc.getNCvar(fname1,'time',timesel=timesel)
#print timedim
if rettype=='dict':
fldret={}
elif rettype=='ndarray':
if zonal:
tmp = cnc.getNCvar(fname1,ncfield,timesel=timesel,calc='zm')*conv
else:
tmp = cnc.getNCvar(fname1,ncfield,timesel=timesel)*conv
if len(tmp.shape)>3:
print '4D variable not supported! @@@@'
return -1
elif len(tmp.shape)==3:
# first dim is time, reshape the 2nd two into one spatial dim
n0=tmp.shape[0]
n1=tmp.shape[1]
n2=tmp.shape[2]
initshape=(len(flist), n0, n1*n2) # 3D, ens x time x space
rshape=initshape[1:] # what to reshape individual data to
elif len(tmp.shape)==2:
initshape=list(tmp.shape)
initshape.insert(0,len(flist))
initshape=tuple(initshape) # 3D, ens x time x space
rshape=None
elif len(tmp.shape)==1:
initshape = (len(flist),tmp.shape[0])
rshape=None
print initshape
fldret = np.zeros(initshape)
print fldret.shape # @@@
else:
print 'return type not supported!! @@@'
return -1
numens = len(flist)
#fldall = np.zeros(numens)
#dlist=[]
for eii,fname in enumerate(flist):
if seas==None:
# don't average, return all months
if zonal:
fld = cnc.getNCvar(fname,ncfield,timesel=timesel,calc='zm')*conv
else:
fld = cnc.getNCvar(fname,ncfield,timesel=timesel)*conv
#if region!=None:
# lat=cnc.getNCvar(fname,'lat')
# lon=cnc.getNCvar(fname,'lon')
# fld = cutl.calc_regmean(fld,lat,lon,region=region)
# do calcs:
if calctype!=None:
print 'calcs not implemented yet!! @@@'
if rettype=='ndarray':
fldseas=fld.reshape(rshape)
else:
if len(fld.shape)!=3:
print 'load LE with rettype dict will break if dim of var is not 3 @@@'
return -1
n0,n1,n2=fld.shape
fldseas=fld.reshape((n0,n1*n2)) # flattening
else:
fldseas={}
for sea in seas:
if zonal:
# @@@ not expecting this to work.
fld = cnc.getNCvar(fname,ncfield,timesel=timesel,seas=sea,calc='zm')*conv
else:
fld = cnc.getNCvar(fname,ncfield,timesel=timesel,seas=sea)*conv
# do calcs
if calctype!=None:
print 'calcs not implemented yet!! @@@'
fldseas[sea]=fld
if rettype=='dict':
fldret[fname]= pd.DataFrame(fldseas,index=timedim)
elif rettype=='ndarray':
if seas!=None:
print 'ndarray return type with a seasonal avg does not work yet @@@@'
fldret[eii,...] = fldseas
#dlist.append(fldseas)
return fldret
def pattcorr_ensemble(ename, field, latlim=60):
# @@@@@@@@@@@@ is this fully implemented? Don't think so. 12/2/14
if ename=='ANT':
ename='HistIC'
elif ename=='TOT':
ename='HistBC'
enssims = con.build_ensemblesims(ename)
ensnum=len(enssims)
# ======= copied from Canam4_BCpatterncorr-Copy0.py ===========
#ensnum=5
diffdict = {}
pcmeandict = {} # fldp-fldc pattern corr compared to mean BC
pchaddict = {} # fldp-fldc pattern corr compared to hadisst
seadiffdict = {} # seasonal mean
pcseameandict = {}
pcsea2meandict = {} # to test the other pattern corr calc
pcsea2pvalmeandict = {} # to test the other pattern corr calc
# generate weights for the pattern corr
lat = con.get_t63lat()
lon = con.get_t63lon()
areas = cutl.calc_cellareas(lat,lon)
areas = areas[lat>latlim,:]
weights = areas / np.sum(np.sum(areas,axis=1),axis=0)
#for eii in range(1,ensnum+1):
for skey in enssims:
#skey = etype + str(eii)
#casenamec = bcasenamec + skey
#casenamep = bcasenamep + skey
#fnamec = basepath + casenamec+ subdir + casenamec + '_' + field + '_001-121_ts.nc'
#fnamep = basepath + casenamep+ subdir + casenamep + '_' + field + '_001-121_ts.nc'
fnamec,fnamep = con.build_filepathpair(skey,field)
# monthly calc
fldc = cnc.getNCvar(fnamec,ncfield,timesel=timesel)*conv
fldp = cnc.getNCvar(fnamep,ncfield,timesel=timesel)*conv
fldd = fldp-fldc
# take the pattern correlation
flddclimo,flddstd = cutl.climatologize(fldd) # climo first (don't need to do for BCs technically)
flddcclimo,flddcstd = cutl.climatologize(flddc) # climo first. baseline diff data
diffdict[skey] = flddclimo
# for each month, compute pattern corr
pc = np.zeros((12))
for mii,mon in enumerate(con.get_mon()):
tmp = np.squeeze(flddclimo[mii,lat>latlim,...])
tmpcmp = np.squeeze(flddcclimo[mii,lat>latlim,...])
pc[mii] = cutl.pattcorr(tmp.flatten()*weights.flatten(),tmpcmp.flatten()*weights.flatten())
pcmeandict[skey] = pc # monthly
# seasonal calc
fldcsea = np.zeros((4,len(lat),len(lon)))
fldpsea = np.zeros((4,len(lat),len(lon)))
flddsea = np.zeros((4,len(lat),len(lon)))
pcsea = np.zeros((4))
pcsea2 = np.zeros((4)) # test pattcorr_pearson() @@
pcsea2pval = np.zeros((4)) # test pattcorr_pearson()
for seaii,sea in enumerate(seasons):
fldcsea[seaii,...] = np.mean(cnc.getNCvar(fnamec,ncfield,timesel=timesel,seas=sea)*conv,axis=0)
fldpsea[seaii,...] = np.mean(cnc.getNCvar(fnamep,ncfield,timesel=timesel,seas=sea)*conv,axis=0)
flddsea[seaii,...] = fldpsea[seaii,...]-fldcsea[seaii,...]
tmp = np.squeeze(flddsea[seaii,lat>latlim,...])
tmpcmp = np.squeeze(flddcsea[seaii,lat>latlim,...])
pcsea[seaii] = cutl.pattcorr(tmp.flatten()*weights.flatten(),
tmpcmp.flatten()*weights.flatten())
pcsea2[seaii],pcsea2pval[seaii] = cutl.pattcorr_pearson(tmp.flatten()*weights.flatten(),
tmpcmp.flatten()*weights.flatten())
seadiffdict[skey] = flddsea
pcseameandict[skey] = pcsea
pcsea2meandict[skey] = pcsea2
pcsea2pvalmeandict[skey] = pcsea2pval
subdir = '/ts/'
##### set up file info and read data #######
if field == 'turb':
field='hfl'; fieldb='hfs'
fnamec = basepath + casename + subdir + casename + '_' + field + '_' + timstr + '_ts.nc'
fnamep1 = basepath + casenamep1 + subdir + casenamep1 + '_' + field + '_' + timstrp + '_ts.nc'
fnamep2 = basepath + casenamep2 + subdir + casenamep2 + '_' + field + '_' + timstrp + '_ts.nc'
fnamep3 = basepath + casenamep3 + subdir + casenamep3 + '_' + field + '_' + timstrp + '_ts.nc'
fnamecb = basepath + casename + subdir + casename + '_' + fieldb + '_' + timstr + '_ts.nc'
fnamep1b = basepath + casenamep1 + subdir + casenamep1 + '_' + fieldb + '_' + timstrp + '_ts.nc'
fnamep2b = basepath + casenamep2 + subdir + casenamep2 + '_' + fieldb + '_' + timstrp + '_ts.nc'
fnamep3b = basepath + casenamep3 + subdir + casenamep3 + '_' + fieldb + '_' + timstrp + '_ts.nc'
fldc = cnc.getNCvar(fnamec,field.upper(),timesel=timesel)*conv +\
cnc.getNCvar(fnamecb,fieldb.upper(),timesel=timesel)*conv
fldp1 = cnc.getNCvar(fnamep1,field.upper(),timesel=timesel)*conv +\
cnc.getNCvar(fnamep1b,fieldb.upper(),timesel=timesel)*conv
fldp2 = cnc.getNCvar(fnamep2,field.upper(),timesel=timesel)*conv +\
cnc.getNCvar(fnamep2b,fieldb.upper(),timesel=timesel)*conv
fldp3 = cnc.getNCvar(fnamep3,field.upper(),timesel=timesel)*conv +\
cnc.getNCvar(fnamep3b,fieldb.upper(),timesel=timesel)*conv
field='turb'
else:
fnamec = basepath + casename + subdir + casename + '_' + field + '_' + timstr + '_ts.nc'
fnamep1 = basepath + casenamep1 + subdir + casenamep1 + '_' + field + '_' + timstrp + '_ts.nc'
fnamep2 = basepath + casenamep2 + subdir + casenamep2 + '_' + field + '_' + timstrp + '_ts.nc'
fnamep3 = basepath + casenamep3 + subdir + casenamep3 + '_' + field + '_' + timstrp + '_ts.nc'
def load_data(fielddict, casename, seas=None, timesel=None,infodict=None,ftype='fullts',
calctype=None, calcdict=None,conv=1, region=None,local=False,
orig=None,verb=True, detrend=True):
""" def load_data(fielddict, seas=('DJF','MAM','JJA','SON'), timesel=None, infodict=None, calctype=None, calcdict=None)
fielddict: should have {'field': field, 'ncfield': ncfield, 'comp': comp} at minimum
ftype: type of filename to build. Right now just 'fullts' for full timeseries
or '1950-2020_climo' or 'ensmean'
seas has to be a tuple
rettype: 'dict' or 'ndarray' as return type.
default is dict of DataFrames (to make a Panel).
else, 'ndarray' is a 4D array
local: is data in ~/work/DATA (local=True) or /raid/ra40
detrend: removes any spurious trend (drift) found in data, for e/ grid point
@@ add regional avg functionality?
@@ region, calctype, calcdict? not implemented
returns an object of type rettype
"""
field = fielddict['field']
ncfield=fielddict['ncfield']
comp = fielddict['comp']
flist = build_filenames(fielddict, casename,ftype=ftype,timesel=timesel,local=local,verb=verb)
for fii,fname in enumerate(flist):
if seas==None:
# don't average, return all months
fld = cnc.getNCvar(fname,ncfield,timesel=timesel)*conv
#if region!=None:
# lat=cnc.getNCvar(fname,'lat')
# lon=cnc.getNCvar(fname,'lon')
# fld = cutl.calc_regmean(fld,lat,lon,region=region)
# do calcs:
if calctype!=None:
print 'calcs not implemented yet!! @@@'
fldseas=fld
else:
sea = seas[0]
fld = cnc.getNCvar(fname,ncfield,timesel=timesel,seas=sea)*conv
# do calcs
if calctype!=None:
print 'calcs not implemented yet!! @@@'
fldseas=fld
# hstack for 1D
if fii==0:
fldret=fldseas
else:
if fldseas.ndim==1:
fldret = np.hstack((fldret,fldseas))
else:
fldret = np.vstack((fldret,fldseas))
if detrend and seas!=None: # @@@ eventually implement for monthly too
flddtr = sp.signal.detrend(fldret,axis=0,type='linear') + fldret.mean(axis=0)
fldret = flddtr
return fldret
fldctmvanNOLON = fldctmvan[:,:,0:-1]
fldctmzmvanNOLON = np.average(fldctmvanNOLON,2)
# vanilla but time-weight
fldczmvan = np.average(fldc,3)
fldczmvanann = cutl.annualize_monthlyts(fldczmvan)
fldczmvananntm = np.average(fldczmvanann,0)
# vanilla but remove extra lon and time-weight
fldcNOLON = fldc[:,:,:,0:-1]
fldczmvanNOLON = np.average(fldcNOLON,3)
fldczmvanannNOLON = cutl.annualize_monthlyts(fldczmvanNOLON)
fldczmvananntmNOLON = np.average(fldczmvanannNOLON,0)
################################# PY/NC ########################################
lat = cnc.getNCvar(fnamec,'lat')
lev = cnc.getNCvar(fnamec,'plev')
#fldczm = cnc.getNCvar(fnamec,ncfield,calc='zm',timechunk=((styr-1)*12,enyr*12+1) )*conv
fldczm = cnc.getNCvar(fnamec,ncfield,calc='zm')*conv
#fldpzm = cnc.getNCvar(fnamep2,ncfield,calc='zm',timechunk=((styr-1)*12,enyr*12+1) )*conv
fldczmtm = np.mean(cutl.annualize_monthlyts(fldczm),0)
fldpzmtm = np.mean(cutl.annualize_monthlyts(fldpzm),0)
fldczmtm2 = np.average(fldczm,0)
fldpzmtm2 = np.average(fldpzm,0) # not month-weighted
################################# CDO ###########################################
## @@ test out cdo bindings and compare
# xm_fgco2 = cdo.zonmean( input = ifile ,returnMaArray ='fgco2' )
# OR: select out specific dates, time-mean then zonal mean
# xm_fgco2 = cdo.zonmean( input = cdo.timmean( input = cdo.seldate('1990-01-01,2005-12-31', input=ifile ) ) ,returnMaArray ='fgco2')
if sicn:
field = 'sicn'
# @@ check what the units are: are they all fraction?
fhadsic = basepath + 'HadISST/hadisst1.1_bc_128_64_1870_2013m03_sicn_' + timeper + 'climo.nc' #SICN, 129x64
fhadsicts = basepath2 + 'HadISST/SICN_BC_HadISST_' + timeper + '_1870010100-2011020100.nc' #0000120100-0125120100.nc' # test matlab generated BC file
fhurrsic = basepath + 'HURRELL/MODEL_ICE.T42_' + timeper + 'climo.nc' #SEAICE (%), 128x64
fnsidcsic = basepath + 'NSIDC/nsidc_bt_128x64_1978m11_2011m12_sicn_' + timeper + 'climo.nc' #SICN, 129x64
fhadsicp = basepath + 'HadISST/hadisst1.1_bc_128_64_1870_2013m03_sicn_' + timeperp2 + 'climo.nc' #SICN, 129x64
fhadsicpts = basepath2 + 'HadISST/SICN_BC_HadISST_' + timeperp2 + '_1870010100-2011020100.nc' #0000120100-0125120100.nc' # test matlab generated BC file
fhurrsicp = basepath + 'HURRELL/MODEL_ICE.T42_' + timeperp2 + 'climo.nc' #SEAICE (%), 128x64
fnsidcsicp = basepath + 'NSIDC/nsidc_bt_128x64_1978m11_2011m12_sicn_' + timeperp2 + 'climo.nc' #SICN, 129x64
hadsicc = cnc.getNCvar(fhadsic,'SICN')
hadsiccts = cnc.getNCvar(fhadsicts,'SICN')
hadsiccclimo,junk = cutl.climatologize3d(hadsiccts)
hurrsicc = cnc.getNCvar(fhurrsic,'SEAICE')/100.
hurrtimes = cnc.getNCvar(fhurrsic,'time')
#hurrsicc.resize(hadsicc.shape) # can't resize like this. ??
#np.append(hurrsicc,hurrsicc[:,:,0])#,axis=2) # could not get append to work
#hurrsicc[:,:,len(lon)-1] = hurrsicc[:,:,0] # add a wraparound. nope.
hurrsicc = np.flipud(hurrsicc) # the lats are flipped compared to hadisst and nsidc
nsidcsicc = cnc.getNCvar(fnsidcsic,'SICN')
hadsicp = cnc.getNCvar(fhadsicp,'SICN')
hadsicpts = cnc.getNCvar(fhadsicpts,'SICN')
hadsicpclimo,junk = cutl.climatologize3d(hadsicpts)
## fldctmvanNOLON = fldctmvan[:,:,0:-1]
## fldctmzmvanNOLON = np.average(fldctmvanNOLON,2)
## # vanilla but time-weight
## fldczmvan = np.average(fldc,3)
## fldczmvanann = cutl.annualize_monthlyts(fldczmvan)
## fldczmvananntm = np.average(fldczmvanann,0)
## # vanilla but remove extra lon and time-weight
## fldcNOLON = fldc[:,:,:,0:-1]
## fldczmvanNOLON = np.average(fldcNOLON,3)
## fldczmvanannNOLON = cutl.annualize_monthlyts(fldczmvanNOLON)
## fldczmvananntmNOLON = np.average(fldczmvanannNOLON,0)
################################# PY/NC/CDO ###################################
lat = cnc.getNCvar(fnamec, "lat")
lev = cnc.getNCvar(fnamec, "plev")
nlev = len(lev)
nlat = len(lat)
fldczmnew = np.append(
cnc.getNCvar(fnamec, ncfield, calc="zm") * conv, cnc.getNCvar(fnamec2, ncfield, calc="zm") * conv, axis=0
)
fldczmnewtm = np.mean(cutl.annualize_monthlyts(fldczmnew), 0)
fldczmnewtm2 = np.mean(fldczmnew, 0) # not time-weighted
fldczmnewANN = np.append(
cnc.getNCvar(fnamec, ncfield, calc="zm", seas="ANN") * conv,
cnc.getNCvar(fnamec2, ncfield, calc="zm", seas="ANN") * conv,
axis=0,
