def get_T63_landseamask(shift_lon, mask_antarctica=True, area='land'):
"""
get JSBACH T63 land sea mask
the LS mask is read from the JSBACH init file
area : str
['land','ocean']: When 'land', then the mask returned
is True on land pixels, for ocean it is vice versa.
In any other case, you get a valid field everywhere (globally)
mask_antarctica : bool
if True, then the mask is FALSE over Antarctica (<60S)
"""
ls_file = get_data_pool_directory() \
+ 'data_sources/LSMASK/jsbach_T63_GR15_4tiles_1992.nc'
ls_mask = Data(ls_file, 'slm', read=True, label='T63 land-sea mask',
lat_name='lat', lon_name='lon', shift_lon=shift_lon)
if area == 'land':
msk = ls_mask.data > 0.
elif area == 'ocean':
msk = ls_mask.data == 0.
else:
msk = np.ones(ls_mask.data.shape).astype('bool')
ls_mask.data[~msk] = 0.
ls_mask.data[msk] = 1.
ls_mask.data = ls_mask.data.astype('bool')
if mask_antarctica:
ls_mask.data[ls_mask.lat < -60.] = False
# ensure that also the mask attribute is set properly
ls_mask._apply_mask(~msk)
return ls_mask
def get_T63_landseamask(shift_lon, mask_antarctica=True, area='land'):
"""
get JSBACH T63 land sea mask
the LS mask is read from the JSBACH init file
area : str
['land','ocean']: When 'land', then the mask returned
is True on land pixels, for ocean it is vice versa.
In any other case, you get a valid field everywhere (globally)
mask_antarctica : bool
if True, then the mask is FALSE over Antarctica (<60S)
"""
ls_file = get_data_pool_directory() \
+ 'data_sources/LSMASK/jsbach_T63_GR15_4tiles_1992.nc'
ls_mask = Data(ls_file,
'slm',
read=True,
label='T63 land-sea mask',
lat_name='lat',
lon_name='lon',
shift_lon=shift_lon)
if area == 'land':
msk = ls_mask.data > 0.
elif area == 'ocean':
msk = ls_mask.data == 0.
else:
msk = np.ones(ls_mask.data.shape).astype('bool')
ls_mask.data[~msk] = 0.
ls_mask.data[msk] = 1.
ls_mask.data = ls_mask.data.astype('bool')
if mask_antarctica:
ls_mask.data[ls_mask.lat < -60.] = False
# ensure that also the mask attribute is set properly
ls_mask._apply_mask(~msk)
return ls_mask
def get_model_data_generic(self, interval='season', **kwargs):
"""
unique parameters are:
filename - file basename
variable - name of the variable as the short_name in the netcdf file
kwargs is a dictionary with keys for each model. Then a dictionary with properties follows
"""
if not self.type in kwargs.keys():
print 'WARNING: it is not possible to get data using generic function, as method missing: ', self.type, kwargs.keys()
return None
locdict = kwargs[self.type]
# read settings and details from the keyword arguments
# no defaults; everything should be explicitely specified in either the config file or the dictionaries
varname = locdict.pop('variable')
units = locdict.pop('unit', 'Crazy Unit')
#interval = kwargs.pop('interval') #, 'season') #does not make sense to specifiy a default value as this option is specified by configuration file!
lat_name = locdict.pop('lat_name', 'lat')
lon_name = locdict.pop('lon_name', 'lon')
model_suffix = locdict.pop('model_suffix')
model_prefix = locdict.pop('model_prefix')
file_format = locdict.pop('file_format')
scf = locdict.pop('scale_factor')
valid_mask = locdict.pop('valid_mask')
custom_path = locdict.pop('custom_path', None)
thelevel = locdict.pop('level', None)
target_grid = self._actplot_options['targetgrid']
interpolation = self._actplot_options['interpolation']
if custom_path is None:
filename1 = ("%s%s/merged/%s_%s_%s_%s_%s.%s" %
(self.data_dir, varname, varname, model_prefix, self.model, self.experiment, model_suffix, file_format))
else:
if self.type == 'CMIP5':
filename1 = ("%s/%s_%s_%s_%s_%s.%s" %
(custom_path, varname, model_prefix, self.model, self.experiment, model_suffix, file_format))
elif self.type == 'CMIP5RAW':
filename1 = ("%s/%s_%s_%s_%s_%s.%s" %
(custom_path, varname, model_prefix, self.model, self.experiment, model_suffix, file_format))
elif self.type == 'CMIP5RAWSINGLE':
print 'todo needs implementation!'
assert False
elif self.type == 'CMIP3':
filename1 = ("%s/%s_%s_%s_%s.%s" %
(custom_path, self.experiment, self.model, varname, model_suffix, file_format))
else:
print self.type
raise ValueError('Can not generate filename: invalid model type! %s' % self.type)
force_calc = False
if self.start_time is None:
raise ValueError('Start time needs to be specified')
if self.stop_time is None:
raise ValueError('Stop time needs to be specified')
#/// PREPROCESSING ///
cdo = Cdo()
s_start_time = str(self.start_time)[0:10]
s_stop_time = str(self.stop_time)[0:10]
#1) select timeperiod and generate monthly mean file
if target_grid == 't63grid':
gridtok = 'T63'
else:
gridtok = 'SPECIAL_GRID'
file_monthly = filename1[:-3] + '_' + s_start_time + '_' + s_stop_time + '_' + gridtok + '_monmean.nc' # target filename
file_monthly = get_temporary_directory() + os.path.basename(file_monthly)
sys.stdout.write('\n *** Model file monthly: %s\n' % file_monthly)
if not os.path.exists(filename1):
print 'WARNING: File not existing: ' + filename1
return None
cdo.monmean(options='-f nc', output=file_monthly, input='-' + interpolation + ',' + target_grid + ' -seldate,' + s_start_time + ',' + s_stop_time + ' ' + filename1, force=force_calc)
sys.stdout.write('\n *** Reading model data... \n')
sys.stdout.write(' Interval: ' + interval + '\n')
#2) calculate monthly or seasonal climatology
if interval == 'monthly':
mdata_clim_file = file_monthly[:-3] + '_ymonmean.nc'
mdata_sum_file = file_monthly[:-3] + '_ymonsum.nc'
mdata_N_file = file_monthly[:-3] + '_ymonN.nc'
mdata_clim_std_file = file_monthly[:-3] + '_ymonstd.nc'
cdo.ymonmean(options='-f nc -b 32', output=mdata_clim_file, input=file_monthly, force=force_calc)
cdo.ymonsum(options='-f nc -b 32', output=mdata_sum_file, input=file_monthly, force=force_calc)
cdo.ymonstd(options='-f nc -b 32', output=mdata_clim_std_file, input=file_monthly, force=force_calc)
cdo.div(options='-f nc', output=mdata_N_file, input=mdata_sum_file + ' ' + mdata_clim_file, force=force_calc) # number of samples
elif interval == 'season':
mdata_clim_file = file_monthly[:-3] + '_yseasmean.nc'
mdata_sum_file = file_monthly[:-3] + '_yseassum.nc'
mdata_N_file = file_monthly[:-3] + '_yseasN.nc'
mdata_clim_std_file = file_monthly[:-3] + '_yseasstd.nc'
cdo.yseasmean(options='-f nc -b 32', output=mdata_clim_file, input=file_monthly, force=force_calc)
cdo.yseassum(options='-f nc -b 32', output=mdata_sum_file, input=file_monthly, force=force_calc)
cdo.yseasstd(options='-f nc -b 32', output=mdata_clim_std_file, input=file_monthly, force=force_calc)
cdo.div(options='-f nc -b 32', output=mdata_N_file, input=mdata_sum_file + ' ' + mdata_clim_file, force=force_calc) # number of samples
else:
raise ValueError('Unknown temporal interval. Can not perform preprocessing!')
if not os.path.exists(mdata_clim_file):
return None
#3) read data
if interval == 'monthly':
thetime_cylce = 12
elif interval == 'season':
thetime_cylce = 4
else:
print interval
raise ValueError('Unsupported interval!')
mdata = Data(mdata_clim_file, varname, read=True, label=self._unique_name, unit=units, lat_name=lat_name, lon_name=lon_name, shift_lon=False, scale_factor=scf, level=thelevel, time_cycle=thetime_cylce)
mdata_std = Data(mdata_clim_std_file, varname, read=True, label=self._unique_name + ' std', unit='-', lat_name=lat_name, lon_name=lon_name, shift_lon=False, level=thelevel, time_cycle=thetime_cylce)
mdata.std = mdata_std.data.copy()
del mdata_std
mdata_N = Data(mdata_N_file, varname, read=True, label=self._unique_name + ' std', unit='-', lat_name=lat_name, lon_name=lon_name, shift_lon=False, scale_factor=scf, level=thelevel)
mdata.n = mdata_N.data.copy()
del mdata_N
#ensure that climatology always starts with January, therefore set date and then sort
mdata.adjust_time(year=1700, day=15) # set arbitrary time for climatology
mdata.timsort()
#4) read monthly data
mdata_all = Data(file_monthly, varname, read=True, label=self._unique_name, unit=units, lat_name=lat_name, lon_name=lon_name, shift_lon=False, time_cycle=12, scale_factor=scf, level=thelevel)
mdata_all.adjust_time(day=15)
#mask_antarctica masks everything below 60 degrees S.
#here we only mask Antarctica, if only LAND points shall be used
if valid_mask == 'land':
mask_antarctica = True
elif valid_mask == 'ocean':
mask_antarctica = False
else:
mask_antarctica = False
if target_grid == 't63grid':
mdata._apply_mask(get_T63_landseamask(False, area=valid_mask, mask_antarctica=mask_antarctica))
mdata_all._apply_mask(get_T63_landseamask(False, area=valid_mask, mask_antarctica=mask_antarctica))
else:
tmpmsk = get_generic_landseamask(False, area=valid_mask, target_grid=target_grid, mask_antarctica=mask_antarctica)
mdata._apply_mask(tmpmsk)
mdata_all._apply_mask(tmpmsk)
del tmpmsk
mdata_mean = mdata_all.fldmean()
# return data as a tuple list
retval = (mdata_all.time, mdata_mean, mdata_all)
del mdata_all
return mdata, retval
def xxxxxxxxxxxxxxxxxxxget_surface_shortwave_radiation_down(self, interval='season', force_calc=False, **kwargs):
"""
return data object of
a) seasonal means for SIS
b) global mean timeseries for SIS at original temporal resolution
"""
the_variable = 'rsds'
locdict = kwargs[self.type]
valid_mask = locdict.pop('valid_mask')
if self.start_time is None:
raise ValueError('Start time needs to be specified')
if self.stop_time is None:
raise ValueError('Stop time needs to be specified')
s_start_time = str(self.start_time)[0:10]
s_stop_time = str(self.stop_time)[0:10]
if self.type == 'CMIP5':
filename1 = self.data_dir + 'rsds' + os.sep + self.experiment + '/ready/' + self.model + '/rsds_Amon_' + self.model + '_' + self.experiment + '_ensmean.nc'
elif self.type == 'CMIP5RAW': # raw CMIP5 data based on ensembles
filename1 = self._get_ensemble_filename(the_variable)
elif self.type == 'CMIP5RAWSINGLE':
filename1 = self.get_single_ensemble_file(the_variable, mip='Amon', realm='atmos', temporal_resolution='mon')
else:
raise ValueError('Unknown model type! not supported here!')
if not os.path.exists(filename1):
print ('WARNING file not existing: %s' % filename1)
return None
#/// PREPROCESSING ///
cdo = Cdo()
#1) select timeperiod and generatget_she monthly mean file
file_monthly = filename1[:-3] + '_' + s_start_time + '_' + s_stop_time + '_T63_monmean.nc'
file_monthly = get_temporary_directory() + os.path.basename(file_monthly)
print file_monthly
sys.stdout.write('\n *** Model file monthly: %s\n' % file_monthly)
cdo.monmean(options='-f nc', output=file_monthly, input='-remapcon,t63grid -seldate,' + s_start_time + ',' + s_stop_time + ' ' + filename1, force=force_calc)
sys.stdout.write('\n *** Reading model data... \n')
sys.stdout.write(' Interval: ' + interval + '\n')
#2) calculate monthly or seasonal climatology
if interval == 'monthly':
sis_clim_file = file_monthly[:-3] + '_ymonmean.nc'
sis_sum_file = file_monthly[:-3] + '_ymonsum.nc'
sis_N_file = file_monthly[:-3] + '_ymonN.nc'
sis_clim_std_file = file_monthly[:-3] + '_ymonstd.nc'
cdo.ymonmean(options='-f nc -b 32', output=sis_clim_file, input=file_monthly, force=force_calc)
cdo.ymonsum(options='-f nc -b 32', output=sis_sum_file, input=file_monthly, force=force_calc)
cdo.ymonstd(options='-f nc -b 32', output=sis_clim_std_file, input=file_monthly, force=force_calc)
cdo.div(options='-f nc', output=sis_N_file, input=sis_sum_file + ' ' + sis_clim_file, force=force_calc) # number of samples
elif interval == 'season':
sis_clim_file = file_monthly[:-3] + '_yseasmean.nc'
sis_sum_file = file_monthly[:-3] + '_yseassum.nc'
sis_N_file = file_monthly[:-3] + '_yseasN.nc'
sis_clim_std_file = file_monthly[:-3] + '_yseasstd.nc'
cdo.yseasmean(options='-f nc -b 32', output=sis_clim_file, input=file_monthly, force=force_calc)
cdo.yseassum(options='-f nc -b 32', output=sis_sum_file, input=file_monthly, force=force_calc)
cdo.yseasstd(options='-f nc -b 32', output=sis_clim_std_file, input=file_monthly, force=force_calc)
cdo.div(options='-f nc -b 32', output=sis_N_file, input=sis_sum_file + ' ' + sis_clim_file, force=force_calc) # number of samples
else:
print interval
raise ValueError('Unknown temporal interval. Can not perform preprocessing!')
if not os.path.exists(sis_clim_file):
return None
#3) read data
sis = Data(sis_clim_file, 'rsds', read=True, label=self._unique_name, unit='$W m^{-2}$', lat_name='lat', lon_name='lon', shift_lon=False)
sis_std = Data(sis_clim_std_file, 'rsds', read=True, label=self._unique_name + ' std', unit='-', lat_name='lat', lon_name='lon', shift_lon=False)
sis.std = sis_std.data.copy()
del sis_std
sis_N = Data(sis_N_file, 'rsds', read=True, label=self._unique_name + ' std', unit='-', lat_name='lat', lon_name='lon', shift_lon=False)
sis.n = sis_N.data.copy()
del sis_N
#ensure that climatology always starts with January, therefore set date and then sort
sis.adjust_time(year=1700, day=15) # set arbitrary time for climatology
sis.timsort()
#4) read monthly data
sisall = Data(file_monthly, 'rsds', read=True, label=self._unique_name, unit='W m^{-2}', lat_name='lat', lon_name='lon', shift_lon=False)
if not sisall._is_monthly():
raise ValueError('Timecycle of 12 expected here!')
sisall.adjust_time(day=15)
# land/sea masking ...
if valid_mask == 'land':
mask_antarctica = True
elif valid_mask == 'ocean':
mask_antarctica = False
else:
mask_antarctica = False
sis._apply_mask(get_T63_landseamask(False, mask_antarctica=mask_antarctica, area=valid_mask))
sisall._apply_mask(get_T63_landseamask(False, mask_antarctica=mask_antarctica, area=valid_mask))
sismean = sisall.fldmean()
# return data as a tuple list
retval = (sisall.time, sismean, sisall)
del sisall
# mask areas without radiation (set to invalid): all data < 1 W/m**2
sis.data = np.ma.array(sis.data, mask=sis.data < 1.)
return sis, retval
def get_model_data_generic(self, interval='season', **kwargs):
"""
unique parameters are:
filename - file basename
variable - name of the variable as the short_name in the netcdf file
kwargs is a dictionary with keys for each model. Then a dictionary with properties follows
"""
if not self.type in kwargs.keys():
print ''
print 'WARNING: it is not possible to get data using generic function, as method missing: ', self.type, kwargs.keys(
)
assert False
locdict = kwargs[self.type]
# read settings and details from the keyword arguments
# no defaults; everything should be explicitely specified in either the config file or the dictionaries
varname = locdict.pop('variable', None)
#~ print self.type
#~ print locdict.keys()
assert varname is not None, 'ERROR: provide varname!'
units = locdict.pop('unit', None)
assert units is not None, 'ERROR: provide unit!'
lat_name = locdict.pop('lat_name', 'lat')
lon_name = locdict.pop('lon_name', 'lon')
model_suffix = locdict.pop('model_suffix', None)
model_prefix = locdict.pop('model_prefix', None)
file_format = locdict.pop('file_format')
scf = locdict.pop('scale_factor')
valid_mask = locdict.pop('valid_mask')
custom_path = locdict.pop('custom_path', None)
thelevel = locdict.pop('level', None)
target_grid = self._actplot_options['targetgrid']
interpolation = self._actplot_options['interpolation']
if custom_path is None:
filename1 = self.get_raw_filename(
varname,
**kwargs) # routine needs to be implemented by each subclass
else:
filename1 = custom_path + self.get_raw_filename(varname, **kwargs)
if filename1 is None:
print_log(WARNING, 'No valid model input data')
return None
force_calc = False
if self.start_time is None:
raise ValueError('Start time needs to be specified')
if self.stop_time is None:
raise ValueError('Stop time needs to be specified')
#/// PREPROCESSING ///
cdo = Cdo()
s_start_time = str(self.start_time)[0:10]
s_stop_time = str(self.stop_time)[0:10]
#1) select timeperiod and generate monthly mean file
if target_grid == 't63grid':
gridtok = 'T63'
else:
gridtok = 'SPECIAL_GRID'
file_monthly = filename1[:
-3] + '_' + s_start_time + '_' + s_stop_time + '_' + gridtok + '_monmean.nc' # target filename
file_monthly = get_temporary_directory() + os.path.basename(
file_monthly)
sys.stdout.write('\n *** Model file monthly: %s\n' % file_monthly)
if not os.path.exists(filename1):
print 'WARNING: File not existing: ' + filename1
return None
cdo.monmean(options='-f nc',
output=file_monthly,
input='-' + interpolation + ',' + target_grid +
' -seldate,' + s_start_time + ',' + s_stop_time + ' ' +
filename1,
force=force_calc)
sys.stdout.write('\n *** Reading model data... \n')
sys.stdout.write(' Interval: ' + interval + '\n')
#2) calculate monthly or seasonal climatology
if interval == 'monthly':
mdata_clim_file = file_monthly[:-3] + '_ymonmean.nc'
mdata_sum_file = file_monthly[:-3] + '_ymonsum.nc'
mdata_N_file = file_monthly[:-3] + '_ymonN.nc'
mdata_clim_std_file = file_monthly[:-3] + '_ymonstd.nc'
cdo.ymonmean(options='-f nc -b 32',
output=mdata_clim_file,
input=file_monthly,
force=force_calc)
cdo.ymonsum(options='-f nc -b 32',
output=mdata_sum_file,
input=file_monthly,
force=force_calc)
cdo.ymonstd(options='-f nc -b 32',
output=mdata_clim_std_file,
input=file_monthly,
force=force_calc)
cdo.div(options='-f nc',
output=mdata_N_file,
input=mdata_sum_file + ' ' + mdata_clim_file,
force=force_calc) # number of samples
elif interval == 'season':
mdata_clim_file = file_monthly[:-3] + '_yseasmean.nc'
mdata_sum_file = file_monthly[:-3] + '_yseassum.nc'
mdata_N_file = file_monthly[:-3] + '_yseasN.nc'
mdata_clim_std_file = file_monthly[:-3] + '_yseasstd.nc'
cdo.yseasmean(options='-f nc -b 32',
output=mdata_clim_file,
input=file_monthly,
force=force_calc)
cdo.yseassum(options='-f nc -b 32',
output=mdata_sum_file,
input=file_monthly,
force=force_calc)
cdo.yseasstd(options='-f nc -b 32',
output=mdata_clim_std_file,
input=file_monthly,
force=force_calc)
cdo.div(options='-f nc -b 32',
output=mdata_N_file,
input=mdata_sum_file + ' ' + mdata_clim_file,
force=force_calc) # number of samples
else:
raise ValueError(
'Unknown temporal interval. Can not perform preprocessing!')
if not os.path.exists(mdata_clim_file):
return None
#3) read data
if interval == 'monthly':
thetime_cylce = 12
elif interval == 'season':
thetime_cylce = 4
else:
print interval
raise ValueError('Unsupported interval!')
mdata = Data(mdata_clim_file,
varname,
read=True,
label=self._unique_name,
unit=units,
lat_name=lat_name,
lon_name=lon_name,
shift_lon=False,
scale_factor=scf,
level=thelevel,
time_cycle=thetime_cylce)
mdata_std = Data(mdata_clim_std_file,
varname,
read=True,
label=self._unique_name + ' std',
unit='-',
lat_name=lat_name,
lon_name=lon_name,
shift_lon=False,
level=thelevel,
time_cycle=thetime_cylce)
mdata.std = mdata_std.data.copy()
del mdata_std
mdata_N = Data(mdata_N_file,
varname,
read=True,
label=self._unique_name + ' std',
unit='-',
lat_name=lat_name,
lon_name=lon_name,
shift_lon=False,
scale_factor=scf,
level=thelevel)
mdata.n = mdata_N.data.copy()
del mdata_N
# ensure that climatology always starts with January, therefore set date and then sort
mdata.adjust_time(year=1700,
day=15) # set arbitrary time for climatology
mdata.timsort()
#4) read monthly data
mdata_all = Data(file_monthly,
varname,
read=True,
label=self._unique_name,
unit=units,
lat_name=lat_name,
lon_name=lon_name,
shift_lon=False,
time_cycle=12,
scale_factor=scf,
level=thelevel)
mdata_all.adjust_time(day=15)
#mask_antarctica masks everything below 60 degrees S.
#here we only mask Antarctica, if only LAND points shall be used
if valid_mask == 'land':
mask_antarctica = True
elif valid_mask == 'ocean':
mask_antarctica = False
else:
mask_antarctica = False
if target_grid == 't63grid':
mdata._apply_mask(
get_T63_landseamask(False,
area=valid_mask,
mask_antarctica=mask_antarctica))
mdata_all._apply_mask(
get_T63_landseamask(False,
area=valid_mask,
mask_antarctica=mask_antarctica))
else:
tmpmsk = get_generic_landseamask(False,
area=valid_mask,
target_grid=target_grid,
mask_antarctica=mask_antarctica)
mdata._apply_mask(tmpmsk)
mdata_all._apply_mask(tmpmsk)
del tmpmsk
mdata_mean = mdata_all.fldmean()
mdata._raw_filename = filename1
mdata._monthly_filename = file_monthly
mdata._clim_filename = mdata_clim_file
mdata._varname = varname
# return data as a tuple list
retval = (mdata_all.time, mdata_mean, mdata_all)
del mdata_all
return mdata, retval
def get_generic_landseamask(shift_lon,
mask_antarctica=True,
area='land',
interpolation_method='remapnn',
target_grid='t63grid',
force=False):
"""
get generic land/sea mask. The routine uses the CDO command 'topo'
to generate a 0.5 degree land/sea mask and remaps this
using nearest neighbor
to the target grid
NOTE: using inconsistent land/sea masks between datasets can
result in considerable biases. Note also that
the application of l/s mask is dependent on the spatial resolution
This routine implements a VERY simple approach, but assuming
that all areas >0 m height are land and the rest is ocean.
Parameters
----------
shift_lon : bool
specifies if longitudes shall be shifted
interpolation_method : str
specifies the interpolation method
that shall be used for remapping the 0.5degree data
to the target grid. This can be any of ['remapnn','remapcon',
'remapbil']
target_grid : str
specifies target grid to interpolate to as
similar to CDO remap functions. This can be either a string or
a filename which includes valid geometry information
force : bool
force calculation (removes previous file) = slower
area : str
['land','ocean']. When 'land', then the mask returned
is True on land pixels, for ocean it is vice versa.
in any other case, you get a valid field everywhere
(globally)
mask_antarctica : bool
mask antarctica; if True, then the mask is
FALSE over Antarctice (<60S)
Returns
-------
returns a Data object
"""
print('WARNING: Automatic generation of land/sea mask. \
Ensure that this is what you want!')
cdo = Cdo()
#/// construct output filename.
#If a filename was given for the grid, replace path separators ///
target_grid1 = target_grid.replace(os.sep, '_')
outputfile = get_temporary_directory() + 'land_sea_fractions_' \
+ interpolation_method + '_' + target_grid1 + '.nc'
print 'outfile: ', outputfile
print 'cmd: ', '-remapnn,' + target_grid + ' -topo'
#/// interpolate data to grid using CDO ///
cdo.monmean(options='-f nc',
output=outputfile,
input='-remapnn,' + target_grid + ' -topo',
force=force)
#/// generate L/S mask from topography (land = height > 0.
ls_mask = Data(outputfile,
'topo',
read=True,
label='generic land-sea mask',
lat_name='lat',
lon_name='lon',
shift_lon=shift_lon)
print('Land/sea mask can be found on file: %s' % outputfile)
if area == 'land':
msk = ls_mask.data > 0. # gives land
elif area == 'ocean':
msk = ls_mask.data <= 0.
else:
msk = np.ones(ls_mask.data.shape).astype('bool')
ls_mask.data[~msk] = 0.
ls_mask.data[msk] = 1.
ls_mask.data = ls_mask.data.astype('bool')
#/// mask Antarctica if desired ///
if mask_antarctica:
ls_mask.data[ls_mask.lat < -60.] = False
# ensure that also the mask attribute is set properly
ls_mask._apply_mask(~msk)
return ls_mask
def _do_preprocessing(self, rawfile, varname, s_start_time, s_stop_time, interval='monthly', force_calc=False, valid_mask='global', target_grid='t63grid'):
"""
perform preprocessing
* selection of variable
* temporal subsetting
"""
cdo = Cdo()
if not os.path.exists(rawfile):
print('File not existing! %s ' % rawfile)
return None, None
# calculate monthly means
file_monthly = get_temporary_directory() + os.sep + os.path.basename(rawfile[:-3]) + '_' + varname + '_' + s_start_time + '_' + s_stop_time + '_mm.nc'
if (force_calc) or (not os.path.exists(file_monthly)):
cdo.monmean(options='-f nc', output=file_monthly, input='-seldate,' + s_start_time + ',' + s_stop_time + ' ' + '-selvar,' + varname + ' ' + rawfile, force=force_calc)
else:
pass
if not os.path.exists(file_monthly):
raise ValueError('Monthly preprocessing did not work! %s ' % file_monthly)
# calculate monthly or seasonal climatology
if interval == 'monthly':
mdata_clim_file = file_monthly[:-3] + '_ymonmean.nc'
mdata_sum_file = file_monthly[:-3] + '_ymonsum.nc'
mdata_N_file = file_monthly[:-3] + '_ymonN.nc'
mdata_clim_std_file = file_monthly[:-3] + '_ymonstd.nc'
cdo.ymonmean(options='-f nc -b 32', output=mdata_clim_file, input=file_monthly, force=force_calc)
cdo.ymonsum(options='-f nc -b 32', output=mdata_sum_file, input=file_monthly, force=force_calc)
cdo.ymonstd(options='-f nc -b 32', output=mdata_clim_std_file, input=file_monthly, force=force_calc)
cdo.div(options='-f nc', output=mdata_N_file, input=mdata_sum_file + ' ' + mdata_clim_file, force=force_calc) # number of samples
elif interval == 'season':
mdata_clim_file = file_monthly[:-3] + '_yseasmean.nc'
mdata_sum_file = file_monthly[:-3] + '_yseassum.nc'
mdata_N_file = file_monthly[:-3] + '_yseasN.nc'
mdata_clim_std_file = file_monthly[:-3] + '_yseasstd.nc'
cdo.yseasmean(options='-f nc -b 32', output=mdata_clim_file, input=file_monthly, force=force_calc)
cdo.yseassum(options='-f nc -b 32', output=mdata_sum_file, input=file_monthly, force=force_calc)
cdo.yseasstd(options='-f nc -b 32', output=mdata_clim_std_file, input=file_monthly, force=force_calc)
cdo.div(options='-f nc -b 32', output=mdata_N_file, input=mdata_sum_file + ' ' + mdata_clim_file, force=force_calc) # number of samples
else:
raise ValueError('Unknown temporal interval. Can not perform preprocessing!')
if not os.path.exists(mdata_clim_file):
return None
# read data
if interval == 'monthly':
thetime_cylce = 12
elif interval == 'season':
thetime_cylce = 4
else:
print interval
raise ValueError('Unsupported interval!')
mdata = Data(mdata_clim_file, varname, read=True, label=self.name, shift_lon=False, time_cycle=thetime_cylce, lat_name='lat', lon_name='lon')
mdata_std = Data(mdata_clim_std_file, varname, read=True, label=self.name + ' std', unit='-', shift_lon=False, time_cycle=thetime_cylce, lat_name='lat', lon_name='lon')
mdata.std = mdata_std.data.copy()
del mdata_std
mdata_N = Data(mdata_N_file, varname, read=True, label=self.name + ' std', shift_lon=False, lat_name='lat', lon_name='lon')
mdata.n = mdata_N.data.copy()
del mdata_N
# ensure that climatology always starts with January, therefore set date and then sort
mdata.adjust_time(year=1700, day=15) # set arbitrary time for climatology
mdata.timsort()
#4) read monthly data
mdata_all = Data(file_monthly, varname, read=True, label=self.name, shift_lon=False, time_cycle=12, lat_name='lat', lon_name='lon')
mdata_all.adjust_time(day=15)
#mask_antarctica masks everything below 60 degree S.
#here we only mask Antarctica, if only LAND points shall be used
if valid_mask == 'land':
mask_antarctica = True
elif valid_mask == 'ocean':
mask_antarctica = False
else:
mask_antarctica = False
if target_grid == 't63grid':
mdata._apply_mask(get_T63_landseamask(False, area=valid_mask, mask_antarctica=mask_antarctica))
mdata_all._apply_mask(get_T63_landseamask(False, area=valid_mask, mask_antarctica=mask_antarctica))
else:
tmpmsk = get_generic_landseamask(False, area=valid_mask, target_grid=target_grid, mask_antarctica=mask_antarctica)
mdata._apply_mask(tmpmsk)
mdata_all._apply_mask(tmpmsk)
del tmpmsk
mdata_mean = mdata_all.fldmean()
# return data as a tuple list
retval = (mdata_all.time, mdata_mean, mdata_all)
del mdata_all
return mdata, retval
def get_jsbach_data_generic(self, interval='season', **kwargs):
"""
unique parameters are:
filename - file basename
variable - name of the variable as the short_name in the netcdf file
kwargs is a dictionary with keys for each model. Then a dictionary with properties follows
"""
if not self.type in kwargs.keys():
print 'WARNING: it is not possible to get data using generic function, as method missing: ', self.type, kwargs.keys()
return None
print self.type
print kwargs
locdict = kwargs[self.type]
# read settings and details from the keyword arguments
# no defaults; everything should be explicitely specified in either the config file or the dictionaries
varname = locdict.pop('variable')
units = locdict.pop('unit', 'Unit not specified')
lat_name = locdict.pop('lat_name', 'lat')
lon_name = locdict.pop('lon_name', 'lon')
#model_suffix = locdict.pop('model_suffix')
#model_prefix = locdict.pop('model_prefix')
file_format = locdict.pop('file_format')
scf = locdict.pop('scale_factor')
valid_mask = locdict.pop('valid_mask')
custom_path = locdict.pop('custom_path', None)
thelevel = locdict.pop('level', None)
target_grid = self._actplot_options['targetgrid']
interpolation = self._actplot_options['interpolation']
if self.type != 'JSBACH_RAW2':
print self.type
raise ValueError('Invalid data format here!')
# define from which stream of JSBACH data needs to be taken for specific variables
if varname in ['swdown_acc', 'swdown_reflect_acc']:
filename1 = self.files['jsbach']
elif varname in ['precip_acc']:
filename1 = self.files['land']
elif varname in ['temp2']:
filename1 = self.files['echam']
elif varname in ['var14']: # albedo vis
filename1 = self.files['albedo_vis']
elif varname in ['var15']: # albedo NIR
filename1 = self.files['albedo_nir']
else:
print varname
raise ValueError('Unknown variable type for JSBACH_RAW2 processing!')
force_calc = False
if self.start_time is None:
raise ValueError('Start time needs to be specified')
if self.stop_time is None:
raise ValueError('Stop time needs to be specified')
#/// PREPROCESSING ///
cdo = Cdo()
s_start_time = str(self.start_time)[0:10]
s_stop_time = str(self.stop_time)[0:10]
#1) select timeperiod and generate monthly mean file
if target_grid == 't63grid':
gridtok = 'T63'
else:
gridtok = 'SPECIAL_GRID'
file_monthly = filename1[:-3] + '_' + s_start_time + '_' + s_stop_time + '_' + gridtok + '_monmean.nc' # target filename
file_monthly = get_temporary_directory() + os.path.basename(file_monthly)
sys.stdout.write('\n *** Model file monthly: %s\n' % file_monthly)
if not os.path.exists(filename1):
print 'WARNING: File not existing: ' + filename1
return None
cdo.monmean(options='-f nc', output=file_monthly, input='-' + interpolation + ',' + target_grid + ' -seldate,' + s_start_time + ',' + s_stop_time + ' ' + filename1, force=force_calc)
sys.stdout.write('\n *** Reading model data... \n')
sys.stdout.write(' Interval: ' + interval + '\n')
#2) calculate monthly or seasonal climatology
if interval == 'monthly':
mdata_clim_file = file_monthly[:-3] + '_ymonmean.nc'
mdata_sum_file = file_monthly[:-3] + '_ymonsum.nc'
mdata_N_file = file_monthly[:-3] + '_ymonN.nc'
mdata_clim_std_file = file_monthly[:-3] + '_ymonstd.nc'
cdo.ymonmean(options='-f nc -b 32', output=mdata_clim_file, input=file_monthly, force=force_calc)
cdo.ymonsum(options='-f nc -b 32', output=mdata_sum_file, input=file_monthly, force=force_calc)
cdo.ymonstd(options='-f nc -b 32', output=mdata_clim_std_file, input=file_monthly, force=force_calc)
cdo.div(options='-f nc', output=mdata_N_file, input=mdata_sum_file + ' ' + mdata_clim_file, force=force_calc) # number of samples
elif interval == 'season':
mdata_clim_file = file_monthly[:-3] + '_yseasmean.nc'
mdata_sum_file = file_monthly[:-3] + '_yseassum.nc'
mdata_N_file = file_monthly[:-3] + '_yseasN.nc'
mdata_clim_std_file = file_monthly[:-3] + '_yseasstd.nc'
cdo.yseasmean(options='-f nc -b 32', output=mdata_clim_file, input=file_monthly, force=force_calc)
cdo.yseassum(options='-f nc -b 32', output=mdata_sum_file, input=file_monthly, force=force_calc)
cdo.yseasstd(options='-f nc -b 32', output=mdata_clim_std_file, input=file_monthly, force=force_calc)
cdo.div(options='-f nc -b 32', output=mdata_N_file, input=mdata_sum_file + ' ' + mdata_clim_file, force=force_calc) # number of samples
else:
raise ValueError('Unknown temporal interval. Can not perform preprocessing! ')
if not os.path.exists(mdata_clim_file):
return None
#3) read data
if interval == 'monthly':
thetime_cylce = 12
elif interval == 'season':
thetime_cylce = 4
else:
print interval
raise ValueError('Unsupported interval!')
mdata = Data(mdata_clim_file, varname, read=True, label=self.model, unit=units, lat_name=lat_name, lon_name=lon_name, shift_lon=False, scale_factor=scf, level=thelevel, time_cycle=thetime_cylce)
mdata_std = Data(mdata_clim_std_file, varname, read=True, label=self.model + ' std', unit='-', lat_name=lat_name, lon_name=lon_name, shift_lon=False, level=thelevel, time_cycle=thetime_cylce)
mdata.std = mdata_std.data.copy()
del mdata_std
mdata_N = Data(mdata_N_file, varname, read=True, label=self.model + ' std', unit='-', lat_name=lat_name, lon_name=lon_name, shift_lon=False, scale_factor=scf, level=thelevel)
mdata.n = mdata_N.data.copy()
del mdata_N
#ensure that climatology always starts with J anuary, therefore set date and then sort
mdata.adjust_time(year=1700, day=15) # set arbitrary time for climatology
mdata.timsort()
#4) read monthly data
mdata_all = Data(file_monthly, varname, read=True, label=self.model, unit=units, lat_name=lat_name, lon_name=lon_name, shift_lon=False, time_cycle=12, scale_factor=scf, level=thelevel)
mdata_all.adjust_time(day=15)
if target_grid == 't63grid':
mdata._apply_mask(get_T63_landseamask(False, area=valid_mask))
mdata_all._apply_mask(get_T63_landseamask(False, area=valid_mask))
else:
tmpmsk = get_generic_landseamask(False, area=valid_mask, target_grid=target_grid)
mdata._apply_mask(tmpmsk)
mdata_all._apply_mask(tmpmsk)
del tmpmsk
mdata_mean = mdata_all.fldmean()
# return data as a tuple list
retval = (mdata_all.time, mdata_mean, mdata_all)
del mdata_all
return mdata, retval
filename = download.get_sample_file(name='air', return_object=False)
air = Data(filename, 'air', read=True)
# generate some mask that fits geometry of data. Could be also read from some
# file. We mimic here some irregular mask
mask = np.zeros(air.data[0, :, :].shape)
mask[20:30, 50:60] = 1.
mask[40:60, 100:120] = 1.
mask = np.asarray(mask).astype('bool') # make a bool array
# generate figure for illustration purposes
f = plt.figure()
ax1 = f.add_subplot(2, 2, 1)
ax2 = f.add_subplot(2, 2, 2)
ax3 = f.add_subplot(2, 2, 3)
ax4 = f.add_subplot(2, 2, 4)
f1 = map_plot(air, ax=ax1, title='This is the original data (unprojected)')
ax2.imshow(mask)
ax2.set_title('This is the mask')
# now we apply the mask to the Data object
air._apply_mask(mask) # applies a mask to each timestep
f3 = map_plot(air, ax=ax3, title='Masked data')
# if you want you can estimate automatically the bounding box
air.cut_bounding_box()
f4 = map_plot(air, ax=ax4, title='Cutted data')
plt.show()
def get_generic_landseamask(shift_lon, mask_antarctica=True,
area='land', interpolation_method='remapnn',
target_grid='t63grid', force=False):
"""
get generic land/sea mask. The routine uses the CDO command 'topo'
to generate a 0.5 degree land/sea mask and remaps this
using nearest neighbor
to the target grid
NOTE: using inconsistent land/sea masks between datasets can
result in considerable biases. Note also that
the application of l/s mask is dependent on the spatial resolution
This routine implements a VERY simple approach, but assuming
that all areas >0 m height are land and the rest is ocean.
Parameters
----------
shift_lon : bool
specifies if longitudes shall be shifted
interpolation_method : str
specifies the interpolation method
that shall be used for remapping the 0.5degree data
to the target grid. This can be any of ['remapnn','remapcon',
'remapbil']
target_grid : str
specifies target grid to interpolate to as
similar to CDO remap functions. This can be either a string or
a filename which includes valid geometry information
force : bool
force calculation (removes previous file) = slower
area : str
['land','ocean']. When 'land', then the mask returned
is True on land pixels, for ocean it is vice versa.
in any other case, you get a valid field everywhere
(globally)
mask_antarctica : bool
mask antarctica; if True, then the mask is
FALSE over Antarctice (<60S)
Returns
-------
returns a Data object
"""
print ('WARNING: Automatic generation of land/sea mask. \
Ensure that this is what you want!')
cdo = Cdo()
#/// construct output filename.
#If a filename was given for the grid, replace path separators ///
target_grid1 = target_grid.replace(os.sep, '_')
outputfile = get_temporary_directory() + 'land_sea_fractions_' \
+ interpolation_method + '_' + target_grid1 + '.nc'
print 'outfile: ', outputfile
print 'cmd: ', '-remapnn,' + target_grid + ' -topo'
#/// interpolate data to grid using CDO ///
cdo.monmean(options='-f nc', output=outputfile,
input='-remapnn,' + target_grid + ' -topo', force=force)
#/// generate L/S mask from topography (land = height > 0.
ls_mask = Data(outputfile, 'topo', read=True,
label='generic land-sea mask',
lat_name='lat', lon_name='lon',
shift_lon=shift_lon)
print('Land/sea mask can be found on file: %s' % outputfile)
if area == 'land':
msk = ls_mask.data > 0. # gives land
elif area == 'ocean':
msk = ls_mask.data <= 0.
else:
msk = np.ones(ls_mask.data.shape).astype('bool')
ls_mask.data[~msk] = 0.
ls_mask.data[msk] = 1.
ls_mask.data = ls_mask.data.astype('bool')
#/// mask Antarctica if desired ///
if mask_antarctica:
ls_mask.data[ls_mask.lat < -60.] = False
# ensure that also the mask attribute is set properly
ls_mask._apply_mask(~msk)
return ls_mask
def get_model_data_generic(self, interval="season", **kwargs):
"""
unique parameters are:
filename - file basename
variable - name of the variable as the short_name in the netcdf file
kwargs is a dictionary with keys for each model. Then a dictionary with properties follows
"""
if not self.type in kwargs.keys():
print ""
print "WARNING: it is not possible to get data using generic function, as method missing: ", self.type, kwargs.keys()
assert False
locdict = kwargs[self.type]
# read settings and details from the keyword arguments
# no defaults; everything should be explicitely specified in either the config file or the dictionaries
varname = locdict.pop("variable", None)
# ~ print self.type
# ~ print locdict.keys()
assert varname is not None, "ERROR: provide varname!"
units = locdict.pop("unit", None)
assert units is not None, "ERROR: provide unit!"
lat_name = locdict.pop("lat_name", "lat")
lon_name = locdict.pop("lon_name", "lon")
model_suffix = locdict.pop("model_suffix", None)
model_prefix = locdict.pop("model_prefix", None)
file_format = locdict.pop("file_format")
scf = locdict.pop("scale_factor")
valid_mask = locdict.pop("valid_mask")
custom_path = locdict.pop("custom_path", None)
thelevel = locdict.pop("level", None)
target_grid = self._actplot_options["targetgrid"]
interpolation = self._actplot_options["interpolation"]
if custom_path is None:
filename1 = self.get_raw_filename(varname, **kwargs) # routine needs to be implemented by each subclass
else:
filename1 = custom_path + self.get_raw_filename(varname, **kwargs)
if filename1 is None:
print_log(WARNING, "No valid model input data")
return None
force_calc = False
if self.start_time is None:
raise ValueError("Start time needs to be specified")
if self.stop_time is None:
raise ValueError("Stop time needs to be specified")
# /// PREPROCESSING ///
cdo = Cdo()
s_start_time = str(self.start_time)[0:10]
s_stop_time = str(self.stop_time)[0:10]
# 1) select timeperiod and generate monthly mean file
if target_grid == "t63grid":
gridtok = "T63"
else:
gridtok = "SPECIAL_GRID"
file_monthly = (
filename1[:-3] + "_" + s_start_time + "_" + s_stop_time + "_" + gridtok + "_monmean.nc"
) # target filename
file_monthly = get_temporary_directory() + os.path.basename(file_monthly)
sys.stdout.write("\n *** Model file monthly: %s\n" % file_monthly)
if not os.path.exists(filename1):
print "WARNING: File not existing: " + filename1
return None
cdo.monmean(
options="-f nc",
output=file_monthly,
input="-"
+ interpolation
+ ","
+ target_grid
+ " -seldate,"
+ s_start_time
+ ","
+ s_stop_time
+ " "
+ filename1,
force=force_calc,
)
sys.stdout.write("\n *** Reading model data... \n")
sys.stdout.write(" Interval: " + interval + "\n")
# 2) calculate monthly or seasonal climatology
if interval == "monthly":
mdata_clim_file = file_monthly[:-3] + "_ymonmean.nc"
mdata_sum_file = file_monthly[:-3] + "_ymonsum.nc"
mdata_N_file = file_monthly[:-3] + "_ymonN.nc"
mdata_clim_std_file = file_monthly[:-3] + "_ymonstd.nc"
cdo.ymonmean(options="-f nc -b 32", output=mdata_clim_file, input=file_monthly, force=force_calc)
cdo.ymonsum(options="-f nc -b 32", output=mdata_sum_file, input=file_monthly, force=force_calc)
cdo.ymonstd(options="-f nc -b 32", output=mdata_clim_std_file, input=file_monthly, force=force_calc)
cdo.div(
options="-f nc", output=mdata_N_file, input=mdata_sum_file + " " + mdata_clim_file, force=force_calc
) # number of samples
elif interval == "season":
mdata_clim_file = file_monthly[:-3] + "_yseasmean.nc"
mdata_sum_file = file_monthly[:-3] + "_yseassum.nc"
mdata_N_file = file_monthly[:-3] + "_yseasN.nc"
mdata_clim_std_file = file_monthly[:-3] + "_yseasstd.nc"
cdo.yseasmean(options="-f nc -b 32", output=mdata_clim_file, input=file_monthly, force=force_calc)
cdo.yseassum(options="-f nc -b 32", output=mdata_sum_file, input=file_monthly, force=force_calc)
cdo.yseasstd(options="-f nc -b 32", output=mdata_clim_std_file, input=file_monthly, force=force_calc)
cdo.div(
options="-f nc -b 32",
output=mdata_N_file,
input=mdata_sum_file + " " + mdata_clim_file,
force=force_calc,
) # number of samples
else:
raise ValueError("Unknown temporal interval. Can not perform preprocessing!")
if not os.path.exists(mdata_clim_file):
return None
# 3) read data
if interval == "monthly":
thetime_cylce = 12
elif interval == "season":
thetime_cylce = 4
else:
print interval
raise ValueError("Unsupported interval!")
mdata = Data(
mdata_clim_file,
varname,
read=True,
label=self._unique_name,
unit=units,
lat_name=lat_name,
lon_name=lon_name,
shift_lon=False,
scale_factor=scf,
level=thelevel,
time_cycle=thetime_cylce,
)
mdata_std = Data(
mdata_clim_std_file,
varname,
read=True,
label=self._unique_name + " std",
unit="-",
lat_name=lat_name,
lon_name=lon_name,
shift_lon=False,
level=thelevel,
time_cycle=thetime_cylce,
)
mdata.std = mdata_std.data.copy()
del mdata_std
mdata_N = Data(
mdata_N_file,
varname,
read=True,
label=self._unique_name + " std",
unit="-",
lat_name=lat_name,
lon_name=lon_name,
shift_lon=False,
scale_factor=scf,
level=thelevel,
)
mdata.n = mdata_N.data.copy()
del mdata_N
# ensure that climatology always starts with January, therefore set date and then sort
mdata.adjust_time(year=1700, day=15) # set arbitrary time for climatology
mdata.timsort()
# 4) read monthly data
mdata_all = Data(
file_monthly,
varname,
read=True,
label=self._unique_name,
unit=units,
lat_name=lat_name,
lon_name=lon_name,
shift_lon=False,
time_cycle=12,
scale_factor=scf,
level=thelevel,
)
mdata_all.adjust_time(day=15)
# mask_antarctica masks everything below 60 degrees S.
# here we only mask Antarctica, if only LAND points shall be used
if valid_mask == "land":
mask_antarctica = True
elif valid_mask == "ocean":
mask_antarctica = False
else:
mask_antarctica = False
if target_grid == "t63grid":
mdata._apply_mask(get_T63_landseamask(False, area=valid_mask, mask_antarctica=mask_antarctica))
mdata_all._apply_mask(get_T63_landseamask(False, area=valid_mask, mask_antarctica=mask_antarctica))
else:
tmpmsk = get_generic_landseamask(
False, area=valid_mask, target_grid=target_grid, mask_antarctica=mask_antarctica
)
mdata._apply_mask(tmpmsk)
mdata_all._apply_mask(tmpmsk)
del tmpmsk
mdata_mean = mdata_all.fldmean()
mdata._raw_filename = filename1
mdata._monthly_filename = file_monthly
mdata._clim_filename = mdata_clim_file
mdata._varname = varname
# return data as a tuple list
retval = (mdata_all.time, mdata_mean, mdata_all)
del mdata_all
return mdata, retval
mask[20:30, 50:60] = 1.
mask[40:60, 100:120] = 1.
mask = np.asarray(mask).astype('bool') # make a bool array
# generate figure for illustration purposes
f = plt.figure()
ax1 = f.add_subplot(2,2,1)
ax2 = f.add_subplot(2,2,2)
ax3 = f.add_subplot(2,2,3)
ax4 = f.add_subplot(2,2,4)
f1 = map_plot(air, ax=ax1, title='This is the original data (unprojected)')
ax2.imshow(mask)
ax2.set_title('This is the mask')
# now we apply the mask to the Data object
air._apply_mask(mask) # applies a mask to each timestep
f3 = map_plot(air, ax=ax3, title='Masked data')
# if you want you can estimate automatically the bounding box
air.cut_bounding_box()
f4 = map_plot(air, ax=ax4, title='Cutted data')
plt.show()