-
Notifications
You must be signed in to change notification settings - Fork 28
/
ensemble_run.py
executable file
·342 lines (310 loc) · 14.5 KB
/
ensemble_run.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
#!/usr/bin/env python
import os, sys, csv, time, math, numpy, getpass
from optparse import OptionParser
#Create, run and process a CLM/ALM model ensemble member
# given specified case and parameters (see parm_list and parm_data files)
# Parent case must be pre-built and all namelists in run directory.
# Post-processing calcuates normalized sum of squared errors (SSE) given
# data constraints specified in "constraints" directory"
# DMRicciuto 12/1/2015
#
# Note: This will only work for single-point CLM/ALM compiled with MPI_SERIAL
#-------------------Parse options-----------------------------------------------
parser = OptionParser()
parser.add_option("--runroot", dest="runroot", default="", \
help="Directory where the run would be created")
parser.add_option("--ens_num", dest="ensnum", default=1, \
help="Ensemble member number")
parser.add_option("--parm_list", dest="parm_list", default="", \
help="File containing parameter names/pfts to modify")
parser.add_option("--parm_data", dest="parm_data", default="", \
help="File containing parameter values and ranges")
parser.add_option("--constraints", dest="constraints", default="", \
help="Directory containing constraining variables")
parser.add_option("--norun", dest="norun", default=False, action="store_true", \
help="Don't run model (use for testing purposes)")
parser.add_option("--machine", dest="machine", default="cades", \
help="My machine")
parser.add_option("--casename", dest="casename", default="", \
help = "Name of case to run")
(options, args) = parser.parse_args()
#================= netcdf manipulation functions ===============================#
def getvar(fname, varname):
usescipy = False
try:
import Scientific.IO.NetCDF as netcdf
except ImportError:
import scipy
from scipy.io import netcdf
usescipy = True
if (usescipy):
nffile = netcdf.netcdf_file(fname,"r",mmap=False)
var = nffile.variables[varname]
varvals = var[:].copy() #works for vector only?
nffile.close()
else:
nffile = netcdf.NetCDFFile(fname,"r")
var = nffile.variables[varname]
varvals = var.getValue()
nffile.close()
return varvals
def putvar(fname, varname, varvals):
usescipy = False
try:
import Scientific.IO.NetCDF as netcdf
except ImportError:
import scipy
from scipy.io import netcdf
usescipy = True
if (usescipy):
nffile = netcdf.netcdf_file(fname,"a",mmap=False)
var = nffile.variables[varname]
var[:] = varvals
nffile.close()
else:
nffile = netcdf.NetCDFFile(fname,"a")
var = nffile.variables[varname]
var.assignValue(varvals)
nffile.close()
ierr = 0
return ierr
#======================================================================
UQdir = os.getcwd()
parm_names=[]
parm_indices=[]
parm_values=[]
myinput = open(options.parm_list, 'r')
lnum=0
username = getpass.getuser()
if (options.machine == 'cades' and options.runroot == ''):
options.runroot = '/lustre/pfs1/cades-ccsi/scratch/'+username
elif (options.runroot == ''):
options.runroot = '../../run'
#get parameter names and PFT information
casenames = []
if ('20TR' in options.casename or '1850' in options.casename):
casenames.append(options.casename)
isfullrun = False
else:
casenames.append(options.casename+'_I1850CLM45CBCN_ad_spinup')
casenames.append(options.casename+'_I1850CLM45CBCN')
casenames.append(options.casename+'_I20TRCLM45CBCN')
#for now, hard-code the number of years for ad_spinup and final spinup
nyears_ad_spinup = 250
nyears_final_spinup = 250
isfullrun = True
for s in myinput:
pdata = s.split()
#print pdata
parm_names.append(pdata[0])
parm_indices.append(int(pdata[1]))
myinput.close()
#get parameter values
myinput = open(options.parm_data, 'r')
for s in myinput:
parm_values.append(float(s))
myinput.close()
n_parameters = len(parm_names)
gst=str(100000+int(options.ensnum))
#create ensemble directories from original case(s)
isfirstcase = True
workdir = os.path.abspath('.')
for casename in casenames:
orig_dir = str(os.path.abspath(options.runroot)+'/'+casename+'/run')
ens_dir = os.path.abspath(options.runroot)+'/UQ/'+casename+'/g'+gst[1:]
os.system('mkdir -p '+options.runroot+'/UQ/'+casename+'/g'+gst[1:]+'/timing/checkpoints')
os.system('rm '+ens_dir+'/*.nc')
os.system('cp '+orig_dir+'/*_in* '+ens_dir)
os.system('cp '+orig_dir+'/*nml '+ens_dir)
os.system('cp '+orig_dir+'/*stream* '+ens_dir)
os.system('cp '+orig_dir+'/domain*.nc '+ens_dir)
os.system('cp '+orig_dir+'/surf*.nc '+ens_dir)
#os.system('cp '+orig_dir+'/*.r.*.nc '+ens_dir)
os.system('cp '+orig_dir+'/*.rc '+ens_dir)
os.system('cp '+orig_dir+'/*para*.nc '+ens_dir)
os.system('cp '+orig_dir+'/*initial* '+ens_dir)
os.system('cp '+orig_dir+'/*pftdyn* '+ens_dir)
username = getpass.getuser()
if (isfullrun): #Full spinup simulation
inifile=''
if ('1850' in casename and 'ad_spinup' not in casename):
yst = str(10000+nyears_ad_spinup+1)
inifile = ens_dir_last+'/'+casename_last+'.clm2.r.'+yst[1:]+'-01-01-00000.nc'
if ('20TR' in casename):
yst = str(10000+nyears_final_spinup+1)
inifile = ens_dir_last+'/'+casename_last+'.clm2.r.'+yst[1:]+'-01-01-00000.nc'
else: #Trasient case only
inifile = ens_dir+'/'+casename+'.clm2.r.1974-01-01-00000.nc'
casename_last = casename
ens_dir_last = ens_dir
#loop through all filenames, change directories in namelists, change parameter values
for f in os.listdir(ens_dir):
if (os.path.isfile(ens_dir+'/'+f) and (f[-2:] == 'in' or f[-3:] == 'nml' or 'streams' in f)):
myinput=open(ens_dir+'/'+f)
myoutput=open(ens_dir+'/'+f+'.tmp','w')
for s in myinput:
if ('paramfile' in s):
est = str(100000+int(options.ensnum))
os.system('cp '+ens_dir+'/clm_param* '+ens_dir+'/clm_params_'+est[1:]+'.nc')
myoutput.write(" paramfile = './clm_params_"+est[1:]+".nc'\n")
#Hard-coded parameter file
pftfile = ens_dir+'/clm_params_'+est[1:]+'.nc'
pnum = 0
for p in parm_names:
#if (pnum == 0):
# stem_leaf = getvar(pftfile, 'stem_leaf')
# stem_leaf[2:5]=-1
# ierr = putvar(pftfile, 'stem_leaf', stem_leaf)
param = getvar(pftfile, p)
if (parm_indices[pnum] > 0):
param[parm_indices[pnum]-1] = parm_values[pnum]
elif (parm_indices[pnum] == 0):
param = parm_values[pnum]
else:
param[:] = parm_values[pnum]
ierr = putvar(pftfile, p, param)
pnum = pnum+1
#elif ('logfile =' in s):
# myoutput.write(s.replace('`date +%y%m%d-%H%M%S`',timestr))
else:
myoutput.write(s.replace(orig_dir,ens_dir))
myoutput.close()
myinput.close()
os.system(' mv '+ens_dir+'/'+f+'.tmp '+ens_dir+'/'+f)
os.chdir(ens_dir)
if (isfirstcase):
exedir = os.path.abspath(orig_dir+'/../bld/')
if (options.norun == False):
if os.path.isfile(exedir+'/acme.exe'):
os.system(exedir+'/acme.exe > acme_log.txt')
elif os.path.isfile(exedir+'/e3sm.exe'):
os.system(exedir+'/e3sm.exe > e3sm_log.txt')
elif os.path.isfile(exedir+'/cesm.exe'):
os.system(exedir+'/cesm.exe > cesm_log.txt')
isfirstcase=False
#--------- code to post-process ensemble member and cacluate total normalized SSE ----------
sse=0
myoutput = open('myoutput_sse.txt','w')
myind = 0
for p in parm_names:
myoutput.write(str(parm_names[myind])+' '+str(parm_indices[myind])+' '+str(parm_values[myind])+'\n')
myind = myind+1
for filename in os.listdir(UQdir+'/'+options.constraints):
if (not os.path.isdir(filename)):
myinput = open(UQdir+'/'+options.constraints+'/'+filename,'r')
myvarname = filename.split('.')[0] #model variable is filename
#code to deal with special variables and/or aggregation
#-------------
lnum = 0
year = 0
for s in myinput:
if (lnum == 0):
header = s.split()
else:
hnum = 0
PFT=-1 #default: don't use PFT-specific info
# if specified, use h1 file (PFT-specific)
doy=-1 #default: annual average
month=-1 #default: don't use monthly data
depth=-1
unc = -999
for h in header:
if (h.lower() == 'year'):
year_info = s.split()[hnum]
if ('-' in year_info):
year_first = int(year_info.split('-')[0])
year_last = int(year_info.split('-')[1])
else:
year_first = int(year_info)
year_last = year_first
if (h.lower() == 'doy'):
doy = int(s.split()[hnum])
if (h.lower() == 'month'):
month = int(s.split()[hnum])
if (h.lower() == 'pft'):
PFT = int(s.split()[hnum])
if (h.lower() == 'value'):
value = float(s.split()[hnum])
if (h.lower() == 'depth'):
depth = float(s.split()[hnum])
if ('unc' in h.lower()):
unc = float(s.split()[hnum])
hnum = hnum+1
#get the relevant variable/dataset
#Assumes annual file(s) with daily output
if (year <= 2016):
if (PFT == -1):
filetype = 'h0'
else:
filetype = 'h1'
file_list = []
for y in range(year_first,year_last+1):
yst_temp = str(10000+y)[1:]
file_list.append(casename+'.clm2.'+filetype+'.'+ \
yst_temp+'-01-01-00000.nc')
#post processing of model output with nco to match constraining variables
for f in file_list:
if (myvarname == 'STEMC'):
os.system('ncap -s "STEMC=DEADSTEMC+LIVESTEMC" '+f+' '+f+'.tmp')
os.system('mv '+f+'.tmp '+f)
if (myvarname == 'AGBIOMASS'):
os.system('ncap -s "AGBIOMASS=DEADSTEMC+LIVESTEMC+LEAFC" '+f+' '+f+'.tmp')
os.system('mv '+f+'.tmp '+f)
isfirstfile = True
for f in file_list:
if (isfirstfile):
myvals = getvar(myfile, myvarname)/(year_last-year_first+1)
isfirstfile = False
else:
myvals = myvals + getvar(myfile, myvarname)/(year_last-year_first+1)
if (doy > 0 and value > -900):
if (PFT > 0):
#PFT-specific constraints
model_val = myvals[doy,PFT-1]
if (unc < 0):
unc = value*0.25 #default uncertainty set to 25%
sse = sse + ((model_val-value) /unc)**2
myoutput.write(str(myvarname)+' '+yst+' '+str(doy)+' '+str(PFT)+' '+ \
str(model_val)+' '+str(value)+' '+str(unc)+' '+str(sse)+'\n')
elif (depth > 0):
#depth-specific constraint (column level only)
layers = [0,1.8,4.5,9.1,16.6,28.9,49.3,82.9,138.3,229.6,343.3]
for l in range(0,10):
if (depth >= layers[l] and depth < layers[l+1]):
thislayer = l
model_val = myvals[doy,thislayer,0]
sse = sse + ((model_val-value) / unc )**2
myoutput.write(str(myvarname)+' '+yst+' '+str(doy)+' '+str(depth)+' '+ \
str(model_val)+' '+str(value)+' '+str(unc)+' '+str(sse)+'\n')
else:
#Daily column-level constraint, no depth/pft information (daily)
model_val = myvals[doy,0]
sse = sse + ((model_val-value) / unc )**2
myoutput.write(str(myvarname)+' '+yst+' '+str(doy)+' '+str(PFT)+' '+ \
str(model_val)+' '+str(value)+' '+str(unc)+' '+str(sse)+'\n')
elif (value > -900):
if (PFT > 0):
#Annual constraints by PFT. Assumes unit conversion from gC/m2/s to gC/m2/yr
model_val = sum(myvals[0:,PFT-1])
if (myvarname == 'NPP' or myvarname == 'GPP' or myvarname == 'NEP' or myvarname == 'NEE'):
model_val = model_val * 24.0 * 3600 #convert to gC/m2/year
else:
model_val = model_val / 365.0 #mean value
sse = sse + ((model_val-value) / unc )**2
myoutput.write(myvarname+' '+yst+' '+str(doy)+' '+str(PFT)+' '+ \
str(model_val)+' '+str(value)+' '+str(unc)+' '+str(sse)+'\n')
else:
#Annual constraints (column level). Assumes unit conversion from gC/m2/s to gC/m2/yr
model_val = sum(myvals[0:,0])
if (myvarname == 'NPP' or myvarname == 'GPP' or myvarname == 'NEP' or myvarname == 'NEE'):
model_val = model_val * 24.0 * 3600 #convert to gC/m2/year
else:
model_val = model_val / 365.0 #mean value
sse = sse + ((model_val-value) / unc )**2
myoutput.write(myvarname+' '+year_info+' '+str(doy)+' '+str(PFT)+' '+ \
str(model_val)+' '+str(value)+' '+str(unc)+' '+str(sse)+'\n')
lnum = lnum+1
myoutput.close()
myoutput = open(workdir+'/qpso_ssedata/mysse_'+gst[1:]+'.txt','w')
myoutput.write(str(sse))
myoutput.close()