def write_chem_species(ncfile, filetimes):
nlev = ncfile.dimensions['lev'].size
# Find all GEOS variables in the categories defined as static variables
b = bpch(filetimes.unique_files()[0])
gc_moz_names = {
} # this will be a dictionary with the GC name as the key and the MOZART-like name as the value
for k in b.variables.keys():
for cat in gc_categories:
if cat in k:
gc_moz_names[k] = geos_to_moz_name(k)
b.close()
for gc_name, moz_name in gc_moz_names.iteritems():
if __debug_level__ > 1:
shell_msg(' Writing {0} as {1}'.format(gc_name, moz_name))
val = []
for fname in filetimes.unique_files():
b = bpch(fname)
this_val = b.variables[gc_name]
this_unit = b.variables[gc_name].units
if this_unit == 'molec/cm3':
if __debug_level__ > 1:
shell_msg('Converting molec/cm3 to VMR')
airden = b.variables['TIME-SER_AIRDEN']
this_val /= airden
else:
if __debug_level__ > 1:
scale = unit_conversion(1.0, 'vmr', this_unit, 'ppp')
shell_msg(' Scaling by {0}'.format(scale))
this_val = unit_conversion(this_val, 'vmr', this_unit, 'ppp')
sz = list(this_val.shape)
n_to_add = nlev - sz[1]
if n_to_add > 0:
sz[1] = n_to_add
padding = np.empty(sz)
padding.fill(fill_val)
this_val = np.concatenate([this_val, padding], 1)
val.append(
flip_dim(this_val, 1)
) # remember, GEOS-Chem defines z=1 as surface; MOZART says that's TOA
b.close()
ncfile.createVariable(moz_name,
np.float32,
dimensions=('time', 'lev', 'lat', 'lon'))
ncfile.variables[moz_name][:] = np.concatenate(val, 0)
ncfile.variables[moz_name].units = 'VMR'
def get_values(wd, species, group, month, debug=False):
print 'Extracting data from the bpch.'
bpch_fname = wd + month + ".ctm.bpch"
print str(bpch_fname)
bpch_data = bpch(bpch_fname)
if debug:
print 'Data groups:'
print bpch_data.groups
print 'Group species:'
print bpch_data.groups[group].variables
species_data = bpch_data.groups[group].variables[species]
# Extract only the data which is in the troposphere.
# This gives us lots of zeroes, which is bad.
time_in_trop = bpch_data.groups['TIME-TPS'].variables['TIMETROP']
species_data = np.multiply(species_data[:, :38, :, :], time_in_trop)
# Trying surface ozone instead of tropospheric ozone
species_data = species_data[0, 0, :, :]
units = bpch_data.groups[group].variables[species].units
start_time = bpch_data.groups[group].variables['tau0']
end_time = bpch_data.groups[group].variables['tau1']
start_time, end_time = extract_month(start_time, end_time)
return species_data, units, start_time, end_time
def get_values( wd, species, group, month, debug=False ):
print 'Extracting data from the bpch.'
bpch_fname = wd + month + ".ctm.bpch"
print str(bpch_fname)
bpch_data = bpch(bpch_fname)
if debug:
print 'Data groups:'
print bpch_data.groups
print 'Group species:'
print bpch_data.groups[group].variables
species_data = bpch_data.groups[group].variables[species]
# Extract only the data which is in the troposphere.
# This gives us lots of zeroes, which is bad.
time_in_trop = bpch_data.groups['TIME-TPS'].variables['TIMETROP']
species_data = np.multiply(species_data[:,:38,:,:], time_in_trop)
# Trying surface ozone instead of tropospheric ozone
species_data = species_data[0,0,:,:]
units = bpch_data.groups[group].variables[species].units
start_time = bpch_data.groups[group].variables['tau0']
end_time = bpch_data.groups[group].variables['tau1']
start_time, end_time = extract_month(start_time, end_time)
return species_data, units, start_time, end_time;
def get_species_data_from_bpch(wd, species, group, debug=True):
print 'Extracting data from the bpch.'
bpch_fname = wd + month + ".ctm.bpch"
# bpch_fname = wd + "ctm.bpch"
bpch_data = bpch(bpch_fname)
if debug:
print 'Data groups:'
print bpch_data.groups
print 'Group species:'
print bpch_data.groups[group].variables
species_data = bpch_data.groups[group].variables[species]
units = bpch_data.groups[group].variables[species].units
start_time = bpch_data.groups[group].variables['tau0']
end_time = bpch_data.groups[group].variables['tau1']
# Turn time from hours from the equinox to YYYY-MM-DD
epoch = datetime.datetime(1985, 1, 1)
start_time = epoch + datetime.timedelta(hours=start_time[0])
end_time = epoch + datetime.timedelta(hours=end_time[0])
# Obtain only the date
start_time = start_time.date()
end_time = end_time.date()
return species_data, units, start_time, end_time
def get_species_data_from_bpch(wd, species, group, debug=True):
print 'Extracting data from the bpch.'
bpch_fname = wd + "ctm.bpch"
# print months
# for month in months:
slice = [2]
tmp_data = bpch(bpch_fname, timeslice=[slice])
if debug:
print 'Data groups:'
print tmp_data.groups
tmp_data_group = tmp_data.groups[group]
if debug:
print 'Group species:'
print tmp_data_group.variables
tmp_data_species = tmp_data_group.variables[species]
tmp_data_units = tmp_data_species.units
print tmp_data_species.shape
data = tmp_data_species
# print range(6)[timeslice]
# combine list into single array
debug = True
if debug:
print 'data length = ' + str(data.dimensions)
# print data
print 'data units = ' + tmp_data_units
# print data
return data
def get_species_data_from_bpch(wd,species,group, debug=True):
print 'Extracting data from the bpch.'
bpch_fname = wd+"ctm.bpch"
# print months
# for month in months:
slice = [2]
tmp_data = bpch(bpch_fname,timeslice = [slice] )
if debug:
print 'Data groups:'
print tmp_data.groups
tmp_data_group = tmp_data.groups[group]
if debug:
print 'Group species:'
print tmp_data_group.variables
tmp_data_species = tmp_data_group.variables[species]
tmp_data_units = tmp_data_species.units
print tmp_data_species.shape
data=tmp_data_species
# print range(6)[timeslice]
# combine list into single array
debug = True
if debug:
print 'data length = ' + str(data.dimensions)
# print data
print 'data units = ' + tmp_data_units
# print data
return data;
def get_species_data_from_bpch(wd,species,group, debug=True):
print 'Extracting data from the bpch.'
bpch_fname = wd + month + ".ctm.bpch"
# bpch_fname = wd + "ctm.bpch"
bpch_data = bpch(bpch_fname)
if debug:
print 'Data groups:'
print bpch_data.groups
print 'Group species:'
print bpch_data.groups[group].variables
species_data = bpch_data.groups[group].variables[species]
units = bpch_data.groups[group].variables[species].units
start_time = bpch_data.groups[group].variables['tau0']
end_time = bpch_data.groups[group].variables['tau1']
# Turn time from hours from the equinox to YYYY-MM-DD
epoch = datetime.datetime(1985,1,1)
start_time = epoch + datetime.timedelta(hours=start_time[0])
end_time = epoch + datetime.timedelta(hours=end_time[0])
# Obtain only the date
start_time = start_time.date()
end_time = end_time.date()
return species_data, units, start_time, end_time;
def write_chem_species(ncfile, filetimes):
nlev = ncfile.dimensions['lev'].size
# Find all GEOS variables in the categories defined as static variables
b = bpch(filetimes.unique_files()[0])
gc_moz_names = {} # this will be a dictionary with the GC name as the key and the MOZART-like name as the value
for k in b.variables.keys():
for cat in gc_categories:
if cat in k:
gc_moz_names[k] = geos_to_moz_name(k)
b.close()
for gc_name, moz_name in gc_moz_names.iteritems():
if __debug_level__ > 1:
shell_msg(' Writing {0} as {1}'.format(gc_name, moz_name))
val = []
for fname in filetimes.unique_files():
b = bpch(fname)
this_val = b.variables[gc_name]
this_unit = b.variables[gc_name].units
if this_unit == 'molec/cm3':
if __debug_level__ > 1:
shell_msg('Converting molec/cm3 to VMR')
airden = b.variables['TIME-SER_AIRDEN']
this_val /= airden
else:
if __debug_level__ > 1:
scale = unit_conversion(1.0, 'vmr', this_unit, 'ppp')
shell_msg(' Scaling by {0}'.format(scale))
this_val = unit_conversion(this_val, 'vmr', this_unit, 'ppp')
sz = list(this_val.shape)
n_to_add = nlev - sz[1]
if n_to_add > 0:
sz[1] = n_to_add
padding = np.empty(sz)
padding.fill(fill_val)
this_val = np.concatenate([this_val, padding], 1)
val.append(flip_dim(this_val, 1)) # remember, GEOS-Chem defines z=1 as surface; MOZART says that's TOA
b.close()
ncfile.createVariable(moz_name, np.float32, dimensions=('time','lev','lat','lon'))
ncfile.variables[moz_name][:] = np.concatenate(val, 0)
ncfile.variables[moz_name].units = 'VMR'
def bpch(fi,group,tracer):
from bpch import bpch
## print 'Using file: '+fi
bc=bpch(fi)
gr=bc.groups[str(group)]
tracer=str(tracer)
data=gr.variables[tracer]
lat=bc.variables['latitude']
lon=bc.variables['longitude']
tau0 = bc.variables['tau0']
layer = bc.variables['layer']
return data, lat, lon, layer, tau0
def populate_times(self, bfiles):
tmp_name_list = []
tmp_date_list = []
for filename in bfiles:
f = bpch(filename)
times = f.variables['time']
for t in times:
tmp_date_list.append(gc_base_date + dt.timedelta(hours=t))
tmp_name_list.append(filename)
date_list = sorted(tmp_date_list)
name_list = []
for d in date_list:
i = tmp_date_list.index(d)
name_list.append(tmp_name_list[i])
return name_list, date_list
def populate_times(self, bfiles):
tmp_name_list = []
tmp_date_list = []
for filename in bfiles:
f = bpch(filename)
times = f.variables['time']
for t in times:
tmp_date_list.append(gc_base_date + dt.timedelta(hours=t))
tmp_name_list.append(filename)
date_list = sorted(tmp_date_list)
name_list = []
for d in date_list:
i = tmp_date_list.index(d)
name_list.append(tmp_name_list[i])
return name_list, date_list
co = np.empty((len(files), len(lat_c), len(lon_c)))
nox = np.empty((len(files), len(lat_c), len(lon_c)))
acet = np.empty((len(files), len(lat_c), len(lon_c)))
isop = np.empty((len(files), len(lat_c), len(lon_c)))
c2h6 = np.empty((len(files), len(lat_c), len(lon_c)))
c3h8 = np.empty((len(files), len(lat_c), len(lon_c)))
ald2 = np.empty((len(files), len(lat_c), len(lon_c)))
alk4 = np.empty((len(files), len(lat_c), len(lon_c)))
ch2o = np.empty((len(files), len(lat_c), len(lon_c)))
mek = np.empty((len(files), len(lat_c), len(lon_c)))
prpe = np.empty((len(files), len(lat_c), len(lon_c)))
for i in range(len(files)):
data = bpch(
files[i],
tracerinfo=
'/work/home/db876/modelling/GEOS_CHEM/v90103/2x2.5/2x2.5_GEOS5_fullgrid/run/valid_tracerinfo.dat'
)
group_pl = data.groups['PORL-L=$']
po3 = group_pl.variables['PO3']
lo3 = group_pl.variables['LO3']
po3 = po3[:, 0, :, :]
lo3 = lo3[:, 0, :, :]
o3_prod[i, :, :] = po3
o3_loss[i, :, :] = lo3
#need to convert voc species from ppbC to ppbV by dividing by number of Carbon atoms in molecule
group_ave = data.groups['IJ-AVG-$']
o3[i, :, :] = group_ave.variables['O3'][:, 0, :, :]
co[i, :, :] = group_ave.variables['CO'][:, 0, :, :]
nox[i, :, :] = group_ave.variables['NOx'][:, 0, :, :]
def define_dimensions(ncfile, filetimes):
"""
Defines all relevant dimensions in the netCDF file and writes any time invariant variables used to describe those
dimensions.
:param ncfile: an instance of netCDF4.Dataset that is the file to be written to
:param filetimes: a FilesAndTimes instance listing all the BPCH files
:return:
"""
if not isinstance(ncfile, ncdf.Dataset):
raise TypeError('ncfile must be an instance of netCDF4.Dataset')
if not isinstance(filetimes, FilesAndTimes):
raise TypeError('filetimes must be an instance of FilesAndTimes')
bfile = bpch(filetimes.files[0])
lon = bfile.variables['longitude']
lon[lon <
0] += 360 # MOZART gives longitude W as positive numbers where -179 -> 181 and -1 -> 359
lat = bfile.variables['latitude']
psurf = bfile.variables['PEDGE-$_PSURF']
# Write the dimensions
ncfile.createDimension('lon', size=len(lon))
ncfile.createDimension('lat', size=len(lat))
ncfile.createDimension('lev', size=len(gc_hya_mid))
ncfile.createDimension('ilev', size=len(gc_hya))
ncfile.createDimension('time', size=0) # unlimited
# and the variables that define their values
ncfile.createVariable('lon', np.float32, dimensions=('lon'))
ncfile.variables['lon'][:] = lon
ncfile.variables['lon'].long_name = 'longitude'
ncfile.variables['lon'].units = 'degrees_east'
ncfile.createVariable('lat', np.float32, dimensions=('lat'))
ncfile.variables['lat'][:] = lat
ncfile.variables['lat'].long_name = 'latitude'
ncfile.variables['lat'].units = 'degrees_north'
# Time dimensions
ncfile.createVariable('time', np.float64, dimensions=('time'))
ncfile.variables['time'].long_name = 'simulation_time'
ncfile.variables['time'].units = 'days since 0000-01-01'
ncfile.variables['time'].calendar = 'gregorian'
ncfile.createVariable('secs', np.int32, dimensions=('time'))
ncfile.variables['secs'].long_name = 'seconds to complete elapsed days'
ncfile.variables['secs'].units = 's'
ncfile.createVariable('date', np.int32, dimensions=('time'))
ncfile.variables[
'date'].long_name = 'current date as 8 digit integer (YYYYMMDD)'
ncfile.createVariable('datesec', np.int32, dimensions=('time'))
ncfile.variables['datesec'].long_name = 'seconds to complete current date'
ncfile.variables['datesec'].units = 's'
date_int, date_sec, days_since, sec_since = mozdate_array(
filetimes.datetimes)
ncfile.variables['date'][:] = date_int
ncfile.variables['datesec'][:] = date_sec
ncfile.variables['time'][:] = days_since
ncfile.variables['secs'][:] = sec_since
# GEOS-Chem uses the hybrid sigma-eta coordinate system
# (http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_vertical_grids#Vertical_grids_for_GEOS-5.2C_GEOS-FP.2C_MERRA.2C_and_MERRA-2)
# and we usually run on the 47-layer reduced vertical grid. The A and B coefficients are given at the above link
# and the formula for converting to actual pressure levels is given at
# http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_vertical_grids#Hybrid_grid_definition
# and is Pedge(I,J,L) = Ap(L) + [ Bp(L) * Psurface(I,J) ], Pcenter(I,J,L) = [ Pedge(I,J,L) + Pedge(I,J,L+1) ]/2
#
# In MOZBC, pressures are computed as ps_mozi(I,J) * hybm + ps0 * hyam; where hybm and hyam are vectors (that are
# the same for every column) of the A and B corefficients at level midpoints, ps_mozi is the MOZART surface pressure
# interpolated to the WRF grid, and ps0 is just a constant, always 100000 Pa. The GEOS-Chem B parameter and the
# MOZART B parameter are both unitless. The GC and MOZART A parameters just differ by a scaling factor, A_moz * P0 =
# A_gc * 100 (GC A is in hPa, moz P0 has units of Pa).
#
# MOZART uses the GMAO standard that the first index in the vertical dimension is the top of the
# atmosphere, whereas GEOS-Chem says the first index is the surface.
ncfile.createVariable('lev', np.float32, dimensions='lev')
ncfile.variables['lev'][:] = 1000 * (gc_hya_mid + gc_hyb_mid)
ncfile.variables[
'lev'].long_name = 'hybrid level at layer midpoints (1000*(A+B))'
ncfile.variables['lev'].units = 'hybrid_sigma_pressure'
ncfile.variables['lev'].positive = 'down'
ncfile.variables['lev'].A_var = 'hyam'
ncfile.variables['lev'].B_var = 'hybm'
ncfile.variables['lev'].P0_var = 'P0'
ncfile.variables['lev'].PS_var = 'PS'
ncfile.variables['lev'].bounds = 'ilev'
ncfile.createVariable('ilev', np.float32, dimensions='ilev')
ncfile.variables['ilev'][:] = 1000 * (gc_hya + gc_hyb)
ncfile.variables[
'ilev'].long_name = 'hybrid level at layer interface (1000*(A+B))'
ncfile.variables['ilev'].units = 'hybrid_sigma_pressure'
ncfile.variables['ilev'].positive = 'down'
ncfile.variables['ilev'].A_var = 'hyai'
ncfile.variables['ilev'].B_var = 'hybi'
ncfile.variables['ilev'].P0_var = 'P0'
ncfile.variables['ilev'].PS_var = 'PS'
ncfile.createVariable('hyam', np.float32, dimensions='lev')
ncfile.variables['hyam'][:] = gc_hya_mid
ncfile.variables[
'hyam'].long_name = 'hybrid A coefficient at layer midpoints'
ncfile.createVariable('hybm', np.float32, dimensions='lev')
ncfile.variables['hybm'][:] = gc_hyb_mid
ncfile.variables[
'hybm'].long_name = 'hybrid B coefficient at layer midpoints'
ncfile.createVariable('hyai', np.float32, dimensions='ilev')
ncfile.variables['hyai'][:] = gc_hya
ncfile.variables[
'hyai'].long_name = 'hybrid A coefficient at layer interfaces'
ncfile.createVariable('hybi', np.float32, dimensions='ilev')
ncfile.variables['hybi'][:] = gc_hyb
ncfile.variables[
'hybi'].long_name = 'hybrid B coefficient at layer interfaces'
ncfile.createVariable('P0', np.float32) # scalar
ncfile.variables['P0'][:] = 1e5
ncfile.variables['P0'].long_name = 'reference pressure'
ncfile.variables['P0'].units = 'Pa'
bfile.close()
# Surface pressure has a time component, so first we need to concatenate all the values
ps_list = []
for fname in filetimes.unique_files():
b = bpch(fname)
ps_list.append(b.variables['PEDGE-$_PSURF'][:, 0, :, :].squeeze() *
100) # get surface only and convert from hPa to Pa
b.close()
ncfile.createVariable('PS', np.float32, dimensions=('time', 'lat', 'lon'))
ncfile.variables['PS'][:] = np.concatenate(ps_list, 0)
ncfile.variables['PS'].units = 'Pa'
import glob
import datetime as datetime
species = 'O3'
#set up plot
fig = plt.figure(figsize=(20, 12))
ax = fig.add_axes([.05, .1, .9, .8])
fig.patch.set_facecolor('white')
#bpch_fname = "bpch_files/4x5_GEOS5/ctm.bpch_4x5_GEOS5_2006-2012"
bpch_fname = "/work/home/bn506/Rate_Sensetivity/Batch_runs/HO2_NO_p0/ctm.bpch"
get_times = [1]
f = bpch(bpch_fname) #,timeslice = get_times)
#lat = f.dimensions['lat']
#lat = f.variables['lat']
#lon = f.variables['lon']
lat = np.arange(-88, 90, 4)
lat = np.insert(lat, 0, -90)
lat = np.append(lat, 90)
lon = np.arange(-182.5, 179, 5)
print lon
tautime = f.variables['tau0']
tautime_end = f.variables['tau1']
def get_species_data_from_bpch(wd,species,group,month, debug=True):
print 'Extracting data from the bpch.'
bpch_fname = wd + month + ".ctm.bpch"
# print months
# for month in months:
bpch_data = bpch(bpch_fname )
if debug:
print 'Data groups:'
print bpch_data.groups
print 'Group species:'
print bpch_data.groups[group].variables
species_data = bpch_data.groups[group].variables[species]
units = bpch_data.groups[group].variables[species].units
start_time = bpch_data.groups[group].variables['tau0']
end_time = bpch_data.groups[group].variables['tau1']
start_time, end_time = extract_month(start_time, end_time)
# #Turn species data from ppbv to moles
#Convert from ppbv to species_moles
air_mass = bpch_data.groups['BXHGHT-$'].variables['AD']
air_moles = (air_mass*1E3)# / ( 0.78*28.0 + 0.22*32.0 )
species_moles = air_moles * species_data
# Get tropopause information
Fraction_of_time_in_the_troposphere = bpch_data.groups["TIME-TPS"].variables['TIMETROP']
if debug:
print Fraction_of_time_in_the_troposphere
# Extract only moles in the troposphere
# by multipling the 4d species concentrations by the fraction of time they were in the troposphere ( 1 = aff trop, 0 = no trop, other = tropopause ).
species_moles = np.multiply(species_moles[:,:38,:,:] , Fraction_of_time_in_the_troposphere)
if debug:
print 'species_moles shape = ' + str(species_moles.shape)
# get the total column moles.
total_column = species_moles[0,:,:,:].sum( axis=0 ) # total_column[t,lat,lon]
if debug:
print' total_column shape = ' + str(total_column.shape)
# Get the surface area
print 'Getting the surface area.'
surface_area = bpch_data.groups['DXYP'].variables['AREA']
if debug:
print 'surface_area shape = ' + str(surface_area.shape)
# surface_area[ lat, long ]
#Convert from total mols to total Volume
conversion_factor = 1E-7#1E-12 * 6.02e23 / 2.69e20
column_volume = np.multiply( total_column, conversion_factor )
#Convert to Dobson units.
dobson_data = np.divide( column_volume , surface_area )
if debug:
print 'dobson_data shape = ' + str(dobson_data.shape)
return dobson_data, units, start_time, end_time;
def pacific_ol(area,dates,pacific_sat_dir,geos_dir,save_location,geosfilestr='ts_satellite_omi'):
#==Options==
#(ref_n,ref_s,ref_e,ref_w) = (40.,0.,180.,160.) #reference area.
(ref_n,ref_s,ref_e,ref_w) = area #reference area.
load_pickle = False #update an existing pickle (True) or create a new one (False)
do_obs = True #work out median obs in reference sector
do_geos = True #work out median model column in reference sector
#dates
first_date = date(2014,1,1)
last_date = date(2014,12,31)
(first_date,last_date) = dates
#get a list of dates
list_of_dates = []
dateclock = first_date
while dateclock <= last_date:
list_of_dates.append(dateclock)
dateclock += td(days=1)
#Load/save loaction
#pickle_location = "/group_workspaces/jasmin/geoschem/local_users/lsurl/CL_PP/bark_2014b.p"
#data locations
#pacific_data_dir = "/group_workspaces/cems2/nceo_generic/nceo_ed/OMHCHO_pacific_bark/"
#geos_dir = "/group_workspaces/jasmin/geoschem/local_users/lsurl/runs/geosfp_2x25_tropchem_2012-2014/"
#==Processing==
#assign dates to files
if do_obs:
files_in_obs_folder = [f for f in listdir(pacific_sat_dir) if isfile(join(pacific_sat_dir, f))] #all files in folder
files_in_obs_folder = [f for f in files_in_obs_folder if f.startswith('OMI-Aura_L2-OMHCHO_')] #only OMI files
obs_file_dates = []
for f in files_in_obs_folder:
this_year = int(f[19:23])
this_month = int(f[24:26])
this_day= int(f[26:28])
obs_file_dates.append(date(this_year,this_month,this_day))
if do_geos:
geosfilestr ='ts_satellite_omi'
print "Looking for ND51 files that begin with '%s'." %geosfilestr
files_in_geos_folder = [f for f in listdir(geos_dir) if isfile(join(geos_dir, f))] #all files in folder
print len(files_in_geos_folder)
#print files_in_geos_folder
files_in_geos_folder = [f for f in files_in_geos_folder if f.startswith(geosfilestr)] #only ND51 files
print len(files_in_geos_folder)
geos_file_dates = []
for f in files_in_geos_folder:
this_year = int(f[len(geosfilestr)+1:len(geosfilestr)+5])
this_month = int(f[len(geosfilestr)+5:len(geosfilestr)+7])
this_day= int(f[len(geosfilestr)+7:len(geosfilestr)+9])
geos_file_dates.append(date(this_year,this_month,this_day))
all_corrections = []
#Cycle through dates
for dateclock in list_of_dates:
today_correction = correction(dateclock)
#observation median
if do_obs:
today_files = [files_in_obs_folder[i] for i in range(0,len(files_in_obs_folder)) if obs_file_dates[i] == dateclock]
today_SC = np.array([])
for f in today_files:
try:
f_HDF = h5py.File(join(pacific_sat_dir, f))
except IOError: #dud files
#print "Cannot open %s" %f
continue
#shorthands
df = f_HDF['HDFEOS']['SWATHS']['OMI Total Column Amount HCHO']['Data Fields']
gf = f_HDF['HDFEOS']['SWATHS']['OMI Total Column Amount HCHO']['Geolocation Fields']
#get data
this_lat = np.array(gf['Latitude']).flatten()
this_lon = np.array(gf['Longitude']).flatten()
this_SC = np.array(df['ColumnAmount']).flatten()
this_XTQF = np.array(gf['XtrackQualityFlags']).flatten()
this_MDQF = np.array(df['MainDataQualityFlag']).flatten()
#geo and quality filter
keep = np.logical_and.reduce((this_lat > ref_s,
this_lat < ref_n,
this_lon > ref_w,
this_lon < ref_e,
this_XTQF == 0,
this_MDQF == 0,
this_SC != np.nan))
this_SC = this_SC[keep]
#print len(this_SC)
today_SC = np.append(today_SC,this_SC)
today_correction.med_obs = np.median(today_SC)
print "Median observed slant column for %s: %g" %(dateclock.strftime("%Y-%m-%d"),today_correction.med_obs)
#geos median
if do_geos:
today_files = [files_in_geos_folder[i] for i in range(0,len(files_in_geos_folder)) if geos_file_dates[i] == dateclock]
if len(today_files) != 1:
today_correction.med_geos = np.nan
print "Median modelled vertical column for %s: nan" %dateclock.strftime("%Y-%m-%d")
else:
f = bpch(join(geos_dir, today_files[0]))
this_hcho = np.array(f.variables["IJ-AVG-$_CH2O"]) * 1e-9 #convert to v/v
box_height = np.array(f.variables['BXHGHT-$_BXHEIGHT']) * 100 #box height in cm
air_den = np.array(f.variables['TIME-SER_AIRDEN']) # air density molec.cm-3
air_amount = np.multiply(box_height,air_den) #air per grid box molec.cm-2
this_data_col = np.sum(np.multiply(this_hcho,air_amount),axis=1) #column molec.cm-2
today_correction.med_geos = np.median(this_data_col)
print "Median modelled vertical column for %s: %g" %(dateclock.strftime("%Y-%m-%d"),today_correction.med_geos)
all_corrections.append(today_correction)
print "saving..."
cPickle.dump(all_corrections,open(save_location,"wb"))
no2_prod = np.empty((len(files), 91, 144))
no2_loss = np.empty((len(files), 91, 144))
ox_prod = np.empty((len(files), 91, 144))
ox_loss = np.empty((len(files), 91, 144))
o3_ave = np.empty((len(files), 91, 144))
no2_ave = np.empty((len(files), 91, 144))
no_ave = np.empty((len(files), 91, 144))
isop_ave = np.empty((len(files), 91, 144))
up_flx = np.empty((len(files), 91, 144))
for i in range(len(files)):
data = bpch(files[i], tracerinfo='valid_tracerinfo.dat')
group_pl = data.groups['PORL-L=$']
group_ave = data.groups['IJ-AVG-$']
group_up = data.groups['UP-FLX-$']
po3 = group_pl.variables['PO3']
print po3.units
lo3 = group_pl.variables['LO3']
pno2 = group_pl.variables['PNO2']
lno2 = group_pl.variables['LNO2']
pox = group_pl.variables['POX']
lox = group_pl.variables['LOX']
o3_a = group_ave.variables['O3']
no_a = group_ave.variables['NOx']
#no2_a = group_ave.variables['NO2']
isop_a = group_ave.variables['ISOP']
def get_species_data_from_bpch(wd, species, group, month, debug=True):
print 'Extracting data from the bpch.'
bpch_fname = wd + month + ".ctm.bpch"
# print months
# for month in months:
bpch_data = bpch(bpch_fname)
if debug:
print 'Data groups:'
print bpch_data.groups
print 'Group species:'
print bpch_data.groups[group].variables
species_data = bpch_data.groups[group].variables[species]
units = bpch_data.groups[group].variables[species].units
start_time = bpch_data.groups[group].variables['tau0']
end_time = bpch_data.groups[group].variables['tau1']
start_time, end_time = extract_month(start_time, end_time)
# #Turn species data from ppbv to moles
#Convert from ppbv to species_moles
air_mass = bpch_data.groups['BXHGHT-$'].variables['AD']
air_moles = (air_mass * 1E3) # / ( 0.78*28.0 + 0.22*32.0 )
species_moles = air_moles * species_data
# Get tropopause information
Fraction_of_time_in_the_troposphere = bpch_data.groups[
"TIME-TPS"].variables['TIMETROP']
if debug:
print Fraction_of_time_in_the_troposphere
# Extract only moles in the troposphere
# by multipling the 4d species concentrations by the fraction of time they were in the troposphere ( 1 = aff trop, 0 = no trop, other = tropopause ).
species_moles = np.multiply(species_moles[:, :38, :, :],
Fraction_of_time_in_the_troposphere)
if debug:
print 'species_moles shape = ' + str(species_moles.shape)
# get the total column moles.
total_column = species_moles[0, :, :, :].sum(
axis=0) # total_column[t,lat,lon]
if debug:
print ' total_column shape = ' + str(total_column.shape)
# Get the surface area
print 'Getting the surface area.'
surface_area = bpch_data.groups['DXYP'].variables['AREA']
if debug:
print 'surface_area shape = ' + str(surface_area.shape)
# surface_area[ lat, long ]
#Convert from total mols to total Volume
conversion_factor = 1E-7 #1E-12 * 6.02e23 / 2.69e20
column_volume = np.multiply(total_column, conversion_factor)
#Convert to Dobson units.
dobson_data = np.divide(column_volume, surface_area)
if debug:
print 'dobson_data shape = ' + str(dobson_data.shape)
return dobson_data, units, start_time, end_time
import glob
import datetime as datetime
species = 'CHBr3'
#set up plot
fig=plt.figure(figsize=(20,12))
ax = fig.add_axes([.05, .1, .9, .8])
fig.patch.set_facecolor('white')
#bpch_fname = "bpch_files/4x5_GEOS5/ctm.bpch_4x5_GEOS5_2006-2012"
bpch_fname = "/home/bn506/GEOS-Run/Budget/Flex/ctm.bpch"
get_times = [4,5,6,8,9]
f = bpch(bpch_fname,timeslice = get_times)
#lat = f.dimensions['lat']
#lat = f.variables['lat']
#lon = f.variables['lon']
lat = np.arange(-88,90,4)
lat = np.insert(lat,0,-90)
lat = np.append(lat,90)
lon = np.arange(-182.5,179,5)
print lon
tautime = f.variables['tau0']
tautime_end = f.variables['tau1']
o3_loss = np.empty((len(files),len(lat_c),len(lon_c)))
o3 = np.empty((len(files),len(lat_c),len(lon_c)))
co = np.empty((len(files),len(lat_c),len(lon_c)))
nox = np.empty((len(files),len(lat_c),len(lon_c)))
acet = np.empty((len(files),len(lat_c),len(lon_c)))
isop = np.empty((len(files),len(lat_c),len(lon_c)))
c2h6 = np.empty((len(files),len(lat_c),len(lon_c)))
c3h8 = np.empty((len(files),len(lat_c),len(lon_c)))
ald2 = np.empty((len(files),len(lat_c),len(lon_c)))
alk4 = np.empty((len(files),len(lat_c),len(lon_c)))
ch2o = np.empty((len(files),len(lat_c),len(lon_c)))
mek = np.empty((len(files),len(lat_c),len(lon_c)))
prpe = np.empty((len(files),len(lat_c),len(lon_c)))
for i in range(len(files)):
data = bpch(files[i],tracerinfo='/work/home/db876/modelling/GEOS_CHEM/v90103/2x2.5/2x2.5_GEOS5_fullgrid/run/valid_tracerinfo.dat')
group_pl = data.groups['PORL-L=$']
po3 = group_pl.variables['PO3']
lo3 = group_pl.variables['LO3']
po3 = po3[:,0,:,:]
lo3 = lo3[:,0,:,:]
o3_prod[i,:,:] = po3
o3_loss[i,:,:] = lo3
#need to convert voc species from ppbC to ppbV by dividing by number of Carbon atoms in molecule
group_ave = data.groups['IJ-AVG-$']
o3[i,:,:] = group_ave.variables['O3'][:,0,:,:]
co[i,:,:] = group_ave.variables['CO'][:,0,:,:]
nox[i,:,:] = group_ave.variables['NOx'][:,0,:,:]
acet[i,:,:] = group_ave.variables['ACET'][:,0,:,:]/3.
isop[i,:,:] = group_ave.variables['ISOP'][:,0,:,:]/5.
def define_dimensions(ncfile, filetimes):
"""
Defines all relevant dimensions in the netCDF file and writes any time invariant variables used to describe those
dimensions.
:param ncfile: an instance of netCDF4.Dataset that is the file to be written to
:param filetimes: a FilesAndTimes instance listing all the BPCH files
:return:
"""
if not isinstance(ncfile, ncdf.Dataset):
raise TypeError('ncfile must be an instance of netCDF4.Dataset')
if not isinstance(filetimes, FilesAndTimes):
raise TypeError('filetimes must be an instance of FilesAndTimes')
bfile = bpch(filetimes.files[0])
lon = bfile.variables['longitude']
lon[lon<0] += 360 # MOZART gives longitude W as positive numbers where -179 -> 181 and -1 -> 359
lat = bfile.variables['latitude']
psurf = bfile.variables['PEDGE-$_PSURF']
# Write the dimensions
ncfile.createDimension('lon', size=len(lon))
ncfile.createDimension('lat', size=len(lat))
ncfile.createDimension('lev', size=len(gc_hya_mid))
ncfile.createDimension('ilev', size=len(gc_hya))
ncfile.createDimension('time', size=0) # unlimited
# and the variables that define their values
ncfile.createVariable('lon', np.float32, dimensions=('lon'))
ncfile.variables['lon'][:] = lon
ncfile.variables['lon'].long_name = 'longitude'
ncfile.variables['lon'].units = 'degrees_east'
ncfile.createVariable('lat', np.float32, dimensions=('lat'))
ncfile.variables['lat'][:] = lat
ncfile.variables['lat'].long_name = 'latitude'
ncfile.variables['lat'].units = 'degrees_north'
# Time dimensions
ncfile.createVariable('time', np.float64, dimensions=('time'))
ncfile.variables['time'].long_name = 'simulation_time'
ncfile.variables['time'].units = 'days since 0000-01-01'
ncfile.variables['time'].calendar = 'gregorian'
ncfile.createVariable('secs', np.int32, dimensions=('time'))
ncfile.variables['secs'].long_name = 'seconds to complete elapsed days'
ncfile.variables['secs'].units = 's'
ncfile.createVariable('date', np.int32, dimensions=('time'))
ncfile.variables['date'].long_name = 'current date as 8 digit integer (YYYYMMDD)'
ncfile.createVariable('datesec', np.int32, dimensions=('time'))
ncfile.variables['datesec'].long_name = 'seconds to complete current date'
ncfile.variables['datesec'].units = 's'
date_int, date_sec, days_since, sec_since = mozdate_array(filetimes.datetimes)
ncfile.variables['date'][:] = date_int
ncfile.variables['datesec'][:] = date_sec
ncfile.variables['time'][:] = days_since
ncfile.variables['secs'][:] = sec_since
# GEOS-Chem uses the hybrid sigma-eta coordinate system
# (http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_vertical_grids#Vertical_grids_for_GEOS-5.2C_GEOS-FP.2C_MERRA.2C_and_MERRA-2)
# and we usually run on the 47-layer reduced vertical grid. The A and B coefficients are given at the above link
# and the formula for converting to actual pressure levels is given at
# http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_vertical_grids#Hybrid_grid_definition
# and is Pedge(I,J,L) = Ap(L) + [ Bp(L) * Psurface(I,J) ], Pcenter(I,J,L) = [ Pedge(I,J,L) + Pedge(I,J,L+1) ]/2
#
# In MOZBC, pressures are computed as ps_mozi(I,J) * hybm + ps0 * hyam; where hybm and hyam are vectors (that are
# the same for every column) of the A and B corefficients at level midpoints, ps_mozi is the MOZART surface pressure
# interpolated to the WRF grid, and ps0 is just a constant, always 100000 Pa. The GEOS-Chem B parameter and the
# MOZART B parameter are both unitless. The GC and MOZART A parameters just differ by a scaling factor, A_moz * P0 =
# A_gc * 100 (GC A is in hPa, moz P0 has units of Pa).
#
# MOZART uses the GMAO standard that the first index in the vertical dimension is the top of the
# atmosphere, whereas GEOS-Chem says the first index is the surface.
ncfile.createVariable('lev', np.float32, dimensions='lev')
ncfile.variables['lev'][:] = 1000*(gc_hya_mid + gc_hyb_mid)
ncfile.variables['lev'].long_name = 'hybrid level at layer midpoints (1000*(A+B))'
ncfile.variables['lev'].units = 'hybrid_sigma_pressure'
ncfile.variables['lev'].positive = 'down'
ncfile.variables['lev'].A_var = 'hyam'
ncfile.variables['lev'].B_var = 'hybm'
ncfile.variables['lev'].P0_var = 'P0'
ncfile.variables['lev'].PS_var = 'PS'
ncfile.variables['lev'].bounds = 'ilev'
ncfile.createVariable('ilev', np.float32, dimensions='ilev')
ncfile.variables['ilev'][:] = 1000*(gc_hya + gc_hyb)
ncfile.variables['ilev'].long_name = 'hybrid level at layer interface (1000*(A+B))'
ncfile.variables['ilev'].units = 'hybrid_sigma_pressure'
ncfile.variables['ilev'].positive = 'down'
ncfile.variables['ilev'].A_var = 'hyai'
ncfile.variables['ilev'].B_var = 'hybi'
ncfile.variables['ilev'].P0_var = 'P0'
ncfile.variables['ilev'].PS_var = 'PS'
ncfile.createVariable('hyam', np.float32, dimensions='lev')
ncfile.variables['hyam'][:] = gc_hya_mid
ncfile.variables['hyam'].long_name = 'hybrid A coefficient at layer midpoints'
ncfile.createVariable('hybm', np.float32, dimensions='lev')
ncfile.variables['hybm'][:] = gc_hyb_mid
ncfile.variables['hybm'].long_name = 'hybrid B coefficient at layer midpoints'
ncfile.createVariable('hyai', np.float32, dimensions='ilev')
ncfile.variables['hyai'][:] = gc_hya
ncfile.variables['hyai'].long_name = 'hybrid A coefficient at layer interfaces'
ncfile.createVariable('hybi', np.float32, dimensions='ilev')
ncfile.variables['hybi'][:] = gc_hyb
ncfile.variables['hybi'].long_name = 'hybrid B coefficient at layer interfaces'
ncfile.createVariable('P0', np.float32) # scalar
ncfile.variables['P0'][:] = 1e5
ncfile.variables['P0'].long_name = 'reference pressure'
ncfile.variables['P0'].units = 'Pa'
bfile.close()
# Surface pressure has a time component, so first we need to concatenate all the values
ps_list = []
for fname in filetimes.unique_files():
b = bpch(fname)
ps_list.append(b.variables['PEDGE-$_PSURF'][:,0,:,:].squeeze()*100) # get surface only and convert from hPa to Pa
b.close()
ncfile.createVariable('PS', np.float32, dimensions=('time', 'lat', 'lon'))
ncfile.variables['PS'][:] = np.concatenate(ps_list, 0)
ncfile.variables['PS'].units = 'Pa'
