-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathoutlier_plot.py
executable file
·288 lines (262 loc) · 12 KB
/
outlier_plot.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
#!/usr/bin/env python
import numpy as np
import numpy.linalg as la
import matplotlib.pyplot as plt
import netCDF4 as nc4
from mg2 import wv_sat_methods as wsm
from mg2 import micro_mg2_0 as mg
HIST_FILE_NAME = "/global/homes/s/santos/project/MG2_data_collection/run/MG2_data_collection.cam.h1.0001-01-06-00000.nc"
file = nc4.Dataset(HIST_FILE_NAME, 'r')
ncol = len(file.dimensions['ncol'])
lev = len(file.dimensions['lev'])
ilev = len(file.dimensions['ilev'])
kind = 8
tmelt = 273.15
h2otrip = 273.16
tboil = 373.16
ttrice = 20.
mwwv = 18.016
mwdair = 28.966
epsilo = mwwv / mwdair
gravit = 9.80616
boltz = 1.38065e-23
avogad = 6.02214e26
rgas = boltz * avogad
rair = rgas / mwdair
rh2o = rgas / mwwv
cpair = 1.00464e3
latvap = 2.501e6
latice = 3.337e5
rhmini = 0.8
dcs = 195.e-6
dcs_tdep = True
uniform = False
do_cldice = True
use_hetfrz_classnuc = True
precip_frac_method = 'in_cloud'
berg_eff_factor = 0.1
allow_sed_supersat = False
ice_sed_ai = 500.
prc_coef1 = 30500.
prc_exp = 3.19
prc_exp1 = -1.2
cld_sed = 1.
mg_prc_coeff_fix = True
errstring = wsm.wv_sat_methods_init(kind, tmelt, h2otrip, tboil, ttrice, epsilo)
if str(errstring).strip() != '':
print("wv_sat_methods initialization error: ", errstring)
errstring = mg.micro_mg_init(kind, gravit, rair, rh2o, cpair, tmelt, latvap,
latice, rhmini, dcs, dcs_tdep, uniform, do_cldice,
use_hetfrz_classnuc, precip_frac_method,
berg_eff_factor, allow_sed_supersat, ice_sed_ai,
prc_coef1, prc_exp, prc_exp1, cld_sed,
mg_prc_coeff_fix)
if str(errstring).strip() != '':
print("MG2 initialization error: ", errstring)
mgncol = 128
t = file.variables["MG2IN_T"]
q = file.variables["MG2IN_Q"]
qc = file.variables["MG2IN_QC"]
qi = file.variables["MG2IN_QI"]
nc = file.variables["MG2IN_NC"]
ni = file.variables["MG2IN_NI"]
qr = file.variables["MG2IN_QR"]
qs = file.variables["MG2IN_QS"]
nr = file.variables["MG2IN_NR"]
ns = file.variables["MG2IN_NS"]
relvar = file.variables["MG2IN_RELVAR"]
accre_enhan = file.variables["MG2IN_ACCRE_ENHAN"]
p = file.variables["MG2IN_P"]
pdel = file.variables["MG2IN_PDEL"]
precipf = file.variables["MG2IN_PRECIP"]
liqcldf = file.variables["MG2IN_LIQCLDF"]
icecldf = file.variables["MG2IN_ICECLDF"]
naai = file.variables["MG2IN_NAAI"]
npccn = file.variables["MG2IN_NPCCN"]
rndst = np.empty((t.shape[0], t.shape[1], t.shape[2], 4))
rndst[:,:,:,0] = file.variables["MG2IN_RNDST1"][:]
rndst[:,:,:,1] = file.variables["MG2IN_RNDST2"][:]
rndst[:,:,:,2] = file.variables["MG2IN_RNDST3"][:]
rndst[:,:,:,3] = file.variables["MG2IN_RNDST4"][:]
nacon = np.empty((t.shape[0], t.shape[1], t.shape[2], 4))
nacon[:,:,:,0] = file.variables["MG2IN_NACON1"][:]
nacon[:,:,:,1] = file.variables["MG2IN_NACON2"][:]
nacon[:,:,:,2] = file.variables["MG2IN_NACON3"][:]
nacon[:,:,:,3] = file.variables["MG2IN_NACON4"][:]
frzimm = file.variables["MG2IN_FRZIMM"]
frzcnt = file.variables["MG2IN_FRZCNT"]
frzdep = file.variables["MG2IN_FRZDEP"]
t_loc = np.empty((mgncol, t.shape[1]), order='F')
q_loc = np.empty((mgncol, q.shape[1]), order='F')
qc_loc = np.empty((mgncol, qc.shape[1]), order='F')
qi_loc = np.empty((mgncol, qi.shape[1]), order='F')
nc_loc = np.empty((mgncol, nc.shape[1]), order='F')
ni_loc = np.empty((mgncol, ni.shape[1]), order='F')
qr_loc = np.empty((mgncol, qr.shape[1]), order='F')
qs_loc = np.empty((mgncol, qs.shape[1]), order='F')
nr_loc = np.empty((mgncol, nr.shape[1]), order='F')
ns_loc = np.empty((mgncol, ns.shape[1]), order='F')
relvar_loc = np.empty((mgncol, relvar.shape[1]), order='F')
accre_enhan_loc = np.empty((mgncol, accre_enhan.shape[1]), order='F')
p_loc = np.empty((mgncol, p.shape[1]), order='F')
pdel_loc = np.empty((mgncol, pdel.shape[1]), order='F')
precipf_loc = np.empty((mgncol, precipf.shape[1]), order='F')
liqcldf_loc = np.empty((mgncol, liqcldf.shape[1]), order='F')
icecldf_loc = np.empty((mgncol, icecldf.shape[1]), order='F')
naai_loc = np.empty((mgncol, naai.shape[1]), order='F')
npccn_loc = np.empty((mgncol, npccn.shape[1]), order='F')
rndst_loc = np.empty((mgncol, rndst.shape[1], 4), order='F')
nacon_loc = np.empty((mgncol, nacon.shape[1], 4), order='F')
frzimm_loc = np.empty((mgncol, frzimm.shape[1]), order='F')
frzcnt_loc = np.empty((mgncol, frzcnt.shape[1]), order='F')
frzdep_loc = np.empty((mgncol, frzdep.shape[1]), order='F')
t_in_loc = np.empty((mgncol, t.shape[1]), order='F')
q_in_loc = np.empty((mgncol, q.shape[1]), order='F')
qc_in_loc = np.empty((mgncol, qc.shape[1]), order='F')
qi_in_loc = np.empty((mgncol, qi.shape[1]), order='F')
qr_in_loc = np.empty((mgncol, qr.shape[1]), order='F')
qs_in_loc = np.empty((mgncol, qs.shape[1]), order='F')
total_columns = 2048
final_time = 1800
use_col_num = 2
timesteps = np.array([5, 15, 30, 60, 120, 300, 900])
loc_arrays = {
'T': t_loc,
'Q': q_loc,
'QC': qc_loc,
'QI': qi_loc,
'QR': qr_loc,
'QS': qs_loc,
'T_IN': t_in_loc,
'Q_IN': q_in_loc,
'QC_IN': qc_in_loc,
'QI_IN': qi_in_loc,
'QR_IN': qr_in_loc,
'QS_IN': qs_in_loc,
}
var_names = sorted(list(loc_arrays.keys()))
out_vals = {}
out_fins = {}
finals = {}
final_means = {}
for name in var_names:
out_vals[name] = []
out_fins[name] = []
for i in range(use_col_num):
out_vals[name].append(np.zeros((lev, timesteps.size - 1)))
out_fins[name].append(np.zeros((lev, timesteps.size - 1)))
finals[name] = np.zeros((total_columns, lev))
final_means[name] = np.zeros((lev,))
norm_diffs = [0.]*use_col_num
for it in range(timesteps.size):
assert final_time % timesteps[it] == 0
nsteps = final_time / timesteps[it]
deltat = float(timesteps[it])
for offset in range(total_columns / mgncol):
t_loc[:,:] = t[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
q_loc[:,:] = q[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
qc_loc[:,:] = qc[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
qi_loc[:,:] = qi[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
nc_loc[:,:] = nc[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
ni_loc[:,:] = ni[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
qr_loc[:,:] = qr[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
qs_loc[:,:] = qs[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
nr_loc[:,:] = nr[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
ns_loc[:,:] = ns[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
relvar_loc[:,:] = relvar[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
accre_enhan_loc[:,:] = accre_enhan[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
p_loc[:,:] = p[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
pdel_loc[:,:] = pdel[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
precipf_loc[:,:] = precipf[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
liqcldf_loc[:,:] = liqcldf[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
icecldf_loc[:,:] = icecldf[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
naai_loc[:,:] = naai[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
npccn_loc[:,:] = npccn[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
rndst_loc[:,:,:] = rndst[0,:,offset*mgncol:(offset+1)*mgncol,:].transpose([1, 0, 2])
nacon_loc[:,:,:] = nacon[0,:,offset*mgncol:(offset+1)*mgncol,:].transpose([1, 0, 2])
frzimm_loc[:,:] = frzimm[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
frzcnt_loc[:,:] = frzcnt[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
frzdep_loc[:,:] = frzdep[0,:,offset*mgncol:(offset+1)*mgncol].transpose()
t_in_loc[:,:] = t_loc
q_in_loc[:,:] = q_loc
qc_in_loc[:,:] = qc_loc
qi_in_loc[:,:] = qi_loc
qr_in_loc[:,:] = qr_loc
qs_in_loc[:,:] = qs_loc
for n in range(nsteps):
qcsinksum_rate1ord, tlat, qvlat, qctend, qitend, nctend, nitend, qrtend, \
qstend, nrtend, nstend, effc, effc_fn, effi, prect, preci, nevapr, \
evapsnow, prain, prodsnow, cmeout, deffi, pgamrad, lamcrad, qsout, dsout, \
rflx, sflx, qrout, reff_rain, reff_snow, qcsevap, qisevap, qvres, cmeitot, \
vtrmc, vtrmi, umr, ums, qcsedten, qisedten, qrsedten, qssedten, pratot, \
prctot, mnuccctot, mnuccttot, msacwitot, psacwstot, bergstot, bergtot, \
melttot, homotot, qcrestot, prcitot, praitot, qirestot, mnuccrtot, \
pracstot, meltsdttot, frzrdttot, mnuccdtot, nrout, nsout, refl, arefl, \
areflz, frefl, csrfl, acsrfl, fcsrfl, rercld, ncai, ncal, qrout2, qsout2, \
nrout2, nsout2, drout2, dsout2, freqs, freqr, nfice, qcrat, errstring, \
prer_evap \
= mg.micro_mg_tend(deltat, t_loc, q_loc, qc_loc, qi_loc, nc_loc,
ni_loc, qr_loc, qs_loc, nr_loc, ns_loc,
relvar_loc, accre_enhan_loc, p_loc, pdel_loc,
precipf_loc, liqcldf_loc, icecldf_loc, naai_loc,
npccn_loc, rndst_loc, nacon_loc,
frzimm=frzimm_loc, frzcnt=frzcnt_loc,
frzdep=frzdep_loc, mgncol=mgncol, nlev=lev)
# Should use geopotential!
t_loc += tlat * deltat / cpair
q_loc += qvlat * deltat
q_loc[:,:] = np.where(q_loc < 1.e-12, 1.e-12, q_loc)
qc_loc += qctend * deltat
qc_loc[:,:] = np.where(qc_loc < 0., 0., qc_loc)
qi_loc += qitend * deltat
qi_loc[:,:] = np.where(qi_loc < 0., 0., qi_loc)
qr_loc += qrtend * deltat
qr_loc[:,:] = np.where(qr_loc < 0., 0., qr_loc)
qs_loc += qstend * deltat
qs_loc[:,:] = np.where(qs_loc < 0., 0., qs_loc)
nc_loc += nctend * deltat
nc_loc[:,:] = np.where(nc_loc > 1.e10, 1.e10, np.where(nc_loc < 1.e-12, 1.e-12, nc_loc))
ni_loc += nitend * deltat
ni_loc[:,:] = np.where(nc_loc > 1.e10, 1.e10, np.where(ni_loc < 1.e-12, 1.e-12, ni_loc))
nr_loc += nrtend * deltat
nr_loc[:,:] = np.where(nc_loc > 1.e10, 1.e10, np.where(nr_loc < 1.e-12, 1.e-12, nr_loc))
ns_loc += nstend * deltat
ns_loc[:,:] = np.where(nc_loc > 1.e10, 1.e10, np.where(ns_loc < 1.e-12, 1.e-12, ns_loc))
if it == 0:
for name in var_names:
finals[name][offset*mgncol:(offset+1)*mgncol,:] = loc_arrays[name]
else:
for i in range(mgncol):
norm_diff = la.norm(pdel_loc[i,:] * (finals['Q'][offset*mgncol+i,:] - q_loc[i,:])) / gravit
if norm_diff > norm_diffs[-1]:
norm_diffs.pop()
idx = -1
for j in range(use_col_num-1):
if norm_diff > norm_diffs[j]:
idx = j
break
norm_diffs.insert(idx, norm_diff)
for name in var_names:
for j in range(use_col_num-1, idx, -1):
out_vals[name][j][:,it-1] = out_vals[name][j-1][:,it-1]
out_fins[name][j][:,it-1] = out_fins[name][j-1][:,it-1]
out_vals[name][idx][:,it-1] = loc_arrays[name][i,:]
out_fins[name][idx][:,it-1] = finals[name][offset*mgncol+i,:]
# Do something with final columns.
p_ref = file.variables['lev']
for name in var_names:
final_means[name][:] = finals[name].mean(axis=0)
for it in range(timesteps.size-1):
mean_var = np.zeros((lev,))
for j in range(use_col_num):
mean_var += out_fins[name][j][:,it]
mean_var /= use_col_num
mean_var -= final_means[name]
plt.plot(mean_var, p_ref, label='$\Delta$t={}'.format(timesteps[it+1]))
plt.gca().invert_yaxis()
plt.legend(loc='best')
plt.xlabel('Anomaly in {}'.format(name))
plt.ylabel('Reference pressure (hPa)')
plt.axis('tight')
plt.savefig('./outlier_t5_{}.eps'.format(name))
plt.close()