Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Map variables and streams from MPAS dycore names to MPAS-Analysis names #52

Merged
merged 12 commits into from
Dec 12, 2016
Merged

Map variables and streams from MPAS dycore names to MPAS-Analysis names #52

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
12 commits
Select commit Hold shift + click to select a range
b0123b5
Add to_datetime and to_timedelta to the Date class
xylar Nov 28, 2016
9f2a10c
Make time-series scripts use timeseries_yr1 & 2
xylar Nov 28, 2016
04d4243
Add has_stream and find_stream to streams file reader
xylar Dec 4, 2016
a473604
Add support to mpas_xarray for a variable map
xylar Dec 7, 2016
6bbc655
Add variable map test to mpas_xarray testing
xylar Dec 7, 2016
2162604
Add ocean dicts mapping stream and variable names
xylar Dec 4, 2016
086fa10
Add stream and var maps to run_analysis and to OHC
xylar Dec 7, 2016
3d94d67
Add stream and variable maps to SST analysis
xylar Dec 7, 2016
666978e
Add variable and stream maps to ocean/modelvsobs
xylar Dec 7, 2016
0ed2b5c
Add variable and stream maps to sea-ice analysis
xylar Dec 7, 2016
8de2a50
Update docstrings and formatting in analysis scripts
xylar Dec 8, 2016
9b9434e
Update variable_map --> varmap in analysis scripts and test
xylar Dec 8, 2016
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
183 changes: 108 additions & 75 deletions mpas_analysis/ocean/ocean_modelvsobs.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,8 +3,8 @@
General comparison of 2-d model fields against data. Currently only supports
mixed layer depths (mld) and sea surface temperature (sst)

Author: Luke Van Roekel, Milena Veneziani
Last Modified: 10/24/2016
Author: Luke Van Roekel, Milena Veneziani, Xylar Asay-Davis
Last Modified: 12/06/2016
"""

import matplotlib.pyplot as plt
Expand All @@ -14,22 +14,35 @@
import xarray as xr
import datetime
from netCDF4 import Dataset as netcdf_dataset
import os.path

from ..shared.mpas_xarray.mpas_xarray import preprocess_mpas, remove_repeated_time_index
from ..shared.mpas_xarray.mpas_xarray import preprocess_mpas, \
remove_repeated_time_index
from ..shared.plot.plotting import plot_global_comparison
from ..shared.interpolation.interpolate import interp_fields, init_tree
from ..shared.constants import constants

from ..shared.io import StreamsFile

def ocn_modelvsobs(config, field):

def ocn_modelvsobs(config, field, streamMap=None, variableMap=None):

"""
Plots a comparison of ACME/MPAS output to SST or MLD observations

config is an instance of MpasAnalysisConfigParser containing configuration
options.

field is the name of a field to be analyize (currently one of 'mld' or
'sst')

If present, streamMap is a dictionary of MPAS-O stream names that map to
their mpas_analysis counterparts.

If present, variableMap is a dictionary of MPAS-O variable names that map
to their mpas_analysis counterparts.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Should we have a description of config and field in the doc string? It will be helpful for people who aren't as familiar with the code.

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

A bit of mission creep but I'll do this.

Authors: Luke Van Roekel, Milena Veneziani, Xylar Asay-Davis
Modified: 10/27/2016
Modified: 12/08/2016
"""

# read parameters from config file
Expand All @@ -42,70 +55,62 @@ def ocn_modelvsobs(config, field):
# reading only those that are between the start and end dates
startDate = config.get('time', 'climo_start_date')
endDate = config.get('time', 'climo_end_date')
infiles = streams.readpath('timeSeriesStatsOutput',
startDate=startDate, endDate=endDate)
print 'Reading files {} through {}'.format(infiles[0],infiles[-1])
streamName = streams.find_stream(streamMap['timeSeriesStats'])
infiles = streams.readpath(streamName, startDate=startDate,
endDate=endDate)
print 'Reading files {} through {}'.format(infiles[0], infiles[-1])

plots_dir = config.get('paths', 'plots_dir')
obsdir = config.get('paths', 'obs_' + field.lower() + 'dir')
obsdir = config.get('paths', 'obs_' + field + 'dir')
casename = config.get('case', 'casename')
meshfile = config.get('data', 'mpas_meshfile')
climo_yr1 = config.getint('time', 'climo_yr1')
climo_yr2 = config.getint('time', 'climo_yr2')
yr_offset = config.getint('time', 'yr_offset')

outputTimes = config.getExpression(field.lower() + '_modelvsobs',
outputTimes = config.getExpression(field + '_modelvsobs',
'comparisonTimes')

f = netcdf_dataset(meshfile,mode='r')
f = netcdf_dataset(meshfile, mode='r')
lonCell = f.variables["lonCell"][:]
latCell = f.variables["latCell"][:]

if field.lower() == 'mld':
obs_filename = "%s/holtetalley_mld_climatology.nc" % obsdir
varList = [field]

ds = xr.open_mfdataset(infiles, preprocess=lambda x:
preprocess_mpas(x, yearoffset=yr_offset,
timestr=['xtime_start', 'xtime_end'],
onlyvars=['time_avg_dThreshMLD']))
ds = remove_repeated_time_index(ds)
ds.rename({'time_avg_dThreshMLD':'mpasData'}, inplace = True)
if field == 'mld':

#Load MLD observational data
selvals = None

# Load MLD observational data
obs_filename = "{}/holtetalley_mld_climatology.nc".format(obsdir)
dsData = xr.open_mfdataset(obs_filename)

#Increment month value to be consistent with the model output
# Increment month value to be consistent with the model output
dsData.iMONTH.values += 1

#Rename the time dimension to be consistent with the SST dataset
dsData.rename({'month':'calmonth'}, inplace = True)
dsData.rename({'iMONTH':'month'}, inplace = True)
# Rename the time dimension to be consistent with the SST dataset
dsData.rename({'month': 'calmonth'}, inplace=True)
dsData.rename({'iMONTH': 'month'}, inplace=True)

#rename appropriate observational data for compactness
dsData.rename({'mld_dt_mean':'observationData'}, inplace = True)
obsFieldName = 'mld_dt_mean'

#Reorder dataset for consistence
# Reorder dataset for consistence
dsData = dsData.transpose('month', 'iLON', 'iLAT')

#Set appropriate MLD figure labels
# Set appropriate MLD figure labels
obsTitleLabel = "Observations (HolteTalley density threshold MLD)"
fileOutLabel = "mldHolteTalleyARGO"
unitsLabel = 'm'

elif field.lower() == 'sst':
elif field == 'sst':

ds = xr.open_mfdataset(infiles, preprocess=lambda x:
preprocess_mpas(x, yearoffset=yr_offset,
timestr=['xtime_start', 'xtime_end'],
onlyvars=['time_avg_activeTracers_temperature'],
selvals={'nVertLevels':0}))
ds = remove_repeated_time_index(ds)
ds.rename({'time_avg_activeTracers_temperature':'mpasData'}, inplace = True)
selvals = {'nVertLevels': 0}

obs_filename = "%s/MODEL.SST.HAD187001-198110.OI198111-201203.nc" % obsdir
obs_filename = \
"{}/MODEL.SST.HAD187001-198110.OI198111-201203.nc".format(obsdir)
dsData = xr.open_mfdataset(obs_filename)
#Select years for averaging (pre-industrial or present-day)
#This seems fragile as definitions can change
# Select years for averaging (pre-industrial or present-day)
# This seems fragile as definitions can change
if yr_offset < 1900:
time_start = datetime.datetime(1870, 1, 1)
time_end = datetime.datetime(1900, 12, 31)
Expand All @@ -118,15 +123,25 @@ def ocn_modelvsobs(config, field):
ds_tslice = dsData.sel(time=slice(time_start, time_end))
monthly_clim_data = ds_tslice.groupby('time.month').mean('time')

#Rename the observation data for code compactness
# Rename the observation data for code compactness
dsData = monthly_clim_data.transpose('month', 'lon', 'lat')
dsData.rename({'SST':'observationData'}, inplace = True)
obsFieldName = 'SST'

#Set appropriate figure labels for SST
obsTitleLabel = "Observations (Hadley/OI, %s)" % preIndustrial_txt
# Set appropriate figure labels for SST
obsTitleLabel = \
"Observations (Hadley/OI, {})".format(preIndustrial_txt)
fileOutLabel = "sstHADOI"
unitsLabel = r'$^o$C'

ds = xr.open_mfdataset(
infiles,
preprocess=lambda x: preprocess_mpas(x, yearoffset=yr_offset,
timestr='Time',
onlyvars=varList,
selvals=selvals,
varmap=variableMap))
ds = remove_repeated_time_index(ds)

time_start = datetime.datetime(yr_offset+climo_yr1, 1, 1)
time_end = datetime.datetime(yr_offset+climo_yr2, 12, 31)
ds_tslice = ds.sel(Time=slice(time_start, time_end))
Expand All @@ -136,21 +151,30 @@ def ocn_modelvsobs(config, field):
latData = latData.flatten()
lonData = lonData.flatten()

daysarray = np.ones((12, dsData.observationData.values.shape[1], dsData.observationData.values.shape[2]))
daysarray = np.ones((12, dsData[obsFieldName].values.shape[1],
dsData[obsFieldName].values.shape[2]))

for i, dval in enumerate(constants.dinmonth):
daysarray[i, :, :] = dval
inds = np.where(np.isnan(dsData.observationData[i, :, :].values))
inds = np.where(np.isnan(dsData[obsFieldName][i, :, :].values))
daysarray[i, inds[0], inds[1]] = np.NaN


# initialize interpolation variables
d2, inds2, lonTarg, latTarg = init_tree(np.rad2deg(lonCell), np.rad2deg(latCell), constants.lonmin,
constants.lonmax, constants.latmin, constants.latmax,
constants.dLongitude, constants.dLatitude)
d, inds, lonTargD, latTargD = init_tree(lonData, latData, constants.lonmin, constants.lonmax,
constants.latmin, constants.latmax, constants.dLongitude,
constants.dLatitude)
d2, inds2, lonTarg, latTarg = init_tree(np.rad2deg(lonCell),
np.rad2deg(latCell),
constants.lonmin,
constants.lonmax,
constants.latmin,
constants.latmax,
constants.dLongitude,
constants.dLatitude)
d, inds, lonTargD, latTargD = init_tree(lonData, latData,
constants.lonmin,
constants.lonmax,
constants.latmin,
constants.latmax,
constants.dLongitude,
constants.dLatitude)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Nice! This is much easier to read now.

nLon = lonTarg.shape[0]
nLat = lonTarg.shape[1]

Expand All @@ -163,50 +187,59 @@ def ocn_modelvsobs(config, field):
monthsvalue = constants.monthdictionary[timestring]

if isinstance(monthsvalue, (int, long)):
modeldata = monthly_clim.sel(month=monthsvalue).mpasData.values
obsdata = dsData.sel(month = monthsvalue).observationData.values
modeldata = monthly_clim.sel(month=monthsvalue)[field].values
obsdata = dsData.sel(month=monthsvalue)[obsFieldName].values
else:
modeldata = np.sum(constants.dinmonth[monthsvalue-1]*monthly_clim.sel(month=monthsvalue).
mpasData.values.T, axis=1) / np.sum(constants.dinmonth[monthsvalue-1])
obsdata = np.nansum(daysarray[monthsvalue-1, :, :]*dsData.sel(month = monthsvalue).
observationData.values, axis=0) / np.nansum(daysarray[monthsvalue-1, :, :],
axis=0)

modeldata = (np.sum(
constants.dinmonth[monthsvalue-1] *
monthly_clim.sel(month=monthsvalue)[field].values.T, axis=1) /
np.sum(constants.dinmonth[monthsvalue-1]))
obsdata = (np.nansum(
daysarray[monthsvalue-1, :, :] *
dsData.sel(month=monthsvalue)[obsFieldName].values, axis=0) /
np.nansum(daysarray[monthsvalue-1, :, :], axis=0))

modelOutput[i, :, :] = interp_fields(modeldata, d2, inds2, lonTarg)
observations[i, :, :] = interp_fields(obsdata.flatten(), d, inds, lonTargD)
observations[i, :, :] = interp_fields(obsdata.flatten(), d, inds,
lonTargD)

for i in range(len(outputTimes)):
bias[i, :, :] = modelOutput[i, :, :] - observations[i, :, :]

clevsModelObs = config.getExpression(field.lower() + '_modelvsobs',
clevsModelObs = config.getExpression(field + '_modelvsobs',
'clevsModelObs')
cmap = plt.get_cmap(config.get(field.lower() + '_modelvsobs',
cmap = plt.get_cmap(config.get(field + '_modelvsobs',
'cmapModelObs'))
cmapIndices = config.getExpression(field.lower() + '_modelvsobs',
cmapIndices = config.getExpression(field + '_modelvsobs',
'cmapIndicesModelObs')
cmapModelObs = cols.ListedColormap(cmap(cmapIndices), "cmapModelObs")
clevsDiff = config.getExpression(field.lower() + '_modelvsobs',
clevsDiff = config.getExpression(field + '_modelvsobs',
'clevsDiff')
cmap = plt.get_cmap(config.get(field.lower() + '_modelvsobs', 'cmapDiff'))
cmapIndices = config.getExpression(field.lower() + '_modelvsobs',
cmap = plt.get_cmap(config.get(field + '_modelvsobs', 'cmapDiff'))
cmapIndices = config.getExpression(field + '_modelvsobs',
'cmapIndicesDiff')
cmapDiff = cols.ListedColormap(cmap(cmapIndices), "cmapDiff")

for i in range(len(outputTimes)):
fileout = "{}/{}_{}_{}_years{:04d}-{:04d}.png".format(
plots_dir, fileOutLabel, casename, outputTimes[i], climo_yr1,
climo_yr2)
title = "{} ({}, years {:04d}-{:04d})".format(
field.upper(), outputTimes[i], climo_yr1, climo_yr2)
plot_global_comparison(config,
lonTarg,
latTarg,
modelOutput[i, :, :],
observations[i, :, :],
bias[i, :,:],
bias[i, :, :],
cmapModelObs,
clevsModelObs,
cmapDiff,
clevsDiff,
fileout = "%s/%s_%s_%s_years%04d-%04d.png" % (plots_dir, fileOutLabel,
casename, outputTimes[i], climo_yr1, climo_yr2),
title = "%s (%s, years %04d-%04d)" % (field.upper(), outputTimes[i], climo_yr1, climo_yr2),
modelTitle = "%s" % casename,
obsTitle = obsTitleLabel,
diffTitle = "Model-Observations",
cbarlabel = unitsLabel)
fileout=fileout,
title=title,
modelTitle="{}".format(casename),
obsTitle=obsTitleLabel,
diffTitle="Model-Observations",
cbarlabel=unitsLabel)
Loading