-
Notifications
You must be signed in to change notification settings - Fork 4
/
EnKF_DD_all_time_2DAv2.py
293 lines (209 loc) · 9.08 KB
/
EnKF_DD_all_time_2DAv2.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
'''
@author Ashesh Chattopadhyay
SPENKF with hybrid observations. See paper for more details https://gmd.copernicus.org/preprints/gmd-2021-71/
'''
import numpy as np
import netCDF4 as nc
import scipy.io as sio
file=sio.loadmat('ERA_grid.mat')
lat=file['lat']
lon=file['lon']
#### Testing set is year 2018 #####
fileList_test=[]
fileList_test.append('geopotential_500hPa_2018_5.625deg.nc')
file=nc.Dataset('ERA_Z500_1hour.nc')
Z500=np.asarray(file['input'])
M=np.mean(Z500.flatten())
sdev=np.std(Z500.flatten())
#### This is the experiment with 0.5\sigma_z. Change noise=1 for the second experiment ####
noise=0.5
from matplotlib import pyplot as plt
F=nc.Dataset(fileList_test[0])
Z=np.asarray(F['z'])
TRUTH=Z
[qx,qy]=np.meshgrid(lon,lat)
Z_rs = np.reshape(Z,[np.size(Z,0), int(np.size(Z,1)*np.size(Z,2))])
TRUTH = Z_rs
Z_rs = (Z_rs-M)/sdev
TRUTH = (TRUTH-M)/sdev
for k in range(1,np.size(Z_rs,0)):
Z_rs[k-1,:]=Z_rs[k-1,:]+np.random.normal(0, noise, 2048)
print('length of initial condition',len(Z_rs[0,:]))
def ENKF(x, n, P ,Q, R, obs, model, u_ensemble):
obs=np.reshape(obs,[n,1])
x=np.reshape(x,[n,1])
[U,S,V]=np.linalg.svd(P)
D=np.zeros([n,n])
np.fill_diagonal(D,S)
sqrtP=np.dot(np.dot(U,np.sqrt(D)),U)
ens=np.zeros([n,2*n])
ens[:,0:n]=np.tile(x,(1,n)) + sqrtP
ens[:,n:]=np.tile(x,(1,n)) - sqrtP
## forecasting step,dummy model
for k in range(0, np.size(ens,1)):
u = model.predict(np.reshape(ens[:,k],[1, 32, 64, 1]))
u_ensemble[:,k]=np.reshape(u,(32*64,))
############################
x_prior = np.reshape(np.mean(u_ensemble,1),[n,1])
print('shape pf x_prior',np.shape(x_prior))
print('shape pf obs',np.shape(obs))
cf_ens = ens - np.tile(x_prior,(1,2*n))
P_prior = np.dot(cf_ens,np.transpose(cf_ens))/(2*n - 1)+Q
h_ens = ens
y_prior=np.reshape(np.mean(h_ens,1),[n,1])
ch_ens = h_ens - np.tile(y_prior,(1,2*n))
print('shape pf y_prior',np.shape(y_prior))
P_y = np.dot(ch_ens, np.transpose(ch_ens))/(2*n-1) + R
P_xy = np.dot(cf_ens, np.transpose(ch_ens)) /(2*n-1)
K = np.dot(P_xy,np.linalg.inv(P_y))
P = P_prior - np.dot(np.dot(K,P_y),np.transpose(K))
x = x_prior + np.dot(K,(obs-y_prior))
return x, P
import tensorflow
import keras.backend as K
#from data_manager import ClutteredMNIST
#from visualizer import plot_mnist_sample
#from visualizer import print_evaluation
#from visualizer import plot_mnist_grid
import netCDF4
import numpy as np
from keras.layers import Input, Convolution2D, Convolution1D, MaxPooling2D, Dense, Dropout, \
Flatten, concatenate, Activation, Reshape, \
UpSampling2D,ZeroPadding2D
import keras
from keras.callbacks import History
history = History()
import keras
from keras.layers import Conv2D, Conv2DTranspose, Cropping2D, Concatenate, ZeroPadding2D
from keras.models import load_model
__version__ = 0.1
def CConv2D(filters, kernel_size, strides=(1, 1), activation='linear', padding='valid', kernel_initializer='glorot_uniform', kernel_regularizer=None):
def CConv2D_inner(x):
# padding (see https://www.tensorflow.org/api_guides/python/nn#Convolution)
in_height = int(x.get_shape()[1])
in_width = int(x.get_shape()[2])
if (in_height % strides[0] == 0):
pad_along_height = max(kernel_size[0] - strides[0], 0)
else:
pad_along_height = max(
kernel_size[0] - (in_height % strides[0]), 0)
if (in_width % strides[1] == 0):
pad_along_width = max(kernel_size[1] - strides[1], 0)
else:
pad_along_width = max(kernel_size[1] - (in_width % strides[1]), 0)
pad_top = pad_along_height // 2
pad_bottom = pad_along_height - pad_top
pad_left = pad_along_width // 2
pad_right = pad_along_width - pad_left
# left and right side for padding
pad_left = Cropping2D(cropping=((0, 0), (in_width-pad_left, 0)))(x)
pad_right = Cropping2D(cropping=((0, 0), (0, in_width-pad_right)))(x)
# add padding to incoming image
conc = Concatenate(axis=2)([pad_left, x, pad_right])
# top/bottom padding options
if padding == 'same':
conc = ZeroPadding2D(padding={'top_pad': pad_top,
'bottom_pad': pad_bottom})(conc)
elif padding == 'valid':
pass
else:
raise Exception('Padding "{}" does not exist!'.format(padding))
# perform the circular convolution
cconv2d = Conv2D(filters=filters, kernel_size=kernel_size,
strides=strides, activation=activation,
padding='valid',
kernel_initializer=kernel_initializer,
kernel_regularizer=kernel_regularizer)(conc)
# return circular convolution layer
return cconv2d
return CConv2D_inner
from keras.layers import Input
from keras.models import Model
from keras.layers import Activation
from keras.layers import MaxPool2D
from keras.layers import Flatten
from keras.layers import Conv2D
from keras.layers import Dense
from utils import get_initial_weights
from layers import BilinearInterpolation
def stn(input_shape=(32, 64, 1), sampling_size=(8, 16), num_classes=10):
inputs = Input(shape=input_shape)
conv1 = Convolution2D(32, 5, 5, activation='relu', border_mode='same')(inputs)
conv1 = Convolution2D(32, 5, 5, activation='relu', border_mode='same')(conv1)
pool1 = MaxPooling2D(pool_size=(2, 2))(conv1)
conv2 = Convolution2D(32, 5, 5, activation='relu', border_mode='same')(pool1)
conv2 = Convolution2D(32, 5, 5, activation='relu', border_mode='same')(conv2)
pool2 = MaxPooling2D(pool_size=(2, 2))(conv2)
conv3 = Convolution2D(32, 5, 5, activation='relu', border_mode='same')(pool2)
conv3 = Convolution2D(32, 5, 5, activation='relu', border_mode='same')(conv3)
conv5 = Convolution2D(32, 5, 5, activation='relu', border_mode='same')(conv3)
conv5 = Convolution2D(32, 5, 5, activation='relu', border_mode='same')(conv5)
locnet = Flatten()(conv5)
locnet = Dense(500)(locnet)
locnet = Activation('relu')(locnet)
locnet = Dense(200)(locnet)
locnet = Activation('relu')(locnet)
locnet = Dense(100)(locnet)
locnet = Activation('relu')(locnet)
locnet = Dense(50)(locnet)
locnet = Activation('relu')(locnet)
weights = get_initial_weights(50)
locnet = Dense(6, weights=weights)(locnet)
x = BilinearInterpolation(sampling_size)([inputs, locnet])
up6 = keras.layers.Concatenate(axis=-1)([Convolution2D(32, 2, 2,activation='relu', border_mode='same')(UpSampling2D(size=(2, 2))(x)), conv2])
conv6 = Convolution2D(32, 5, 5, activation='relu', border_mode='same')(up6)
conv6 = Convolution2D(32, 5, 5, activation='relu', border_mode='same')(conv6)
up7 = keras.layers.Concatenate(axis=-1)([Convolution2D(32, 2, 2,activation='relu', border_mode='same')(UpSampling2D(size=(2, 2))(conv6)), conv1])
conv7 = Convolution2D(32, 5, 5, activation='relu', border_mode='same')(up7)
conv7 = Convolution2D(32, 5, 5, activation='relu', border_mode='same')(conv7)
conv10 = Convolution2D(1, 5, 5, activation='linear',border_mode='same')(conv7)
model = Model(input=inputs, output=conv10)
return model
model = stn()
model.load_weights('best_weights_lead1.h5')
###### Start Data Assimilation Process #########################################
model_vir=stn()
model_vir.load_weights('best_weights_lead12.h5')
time = 1200
virtual_obs=np.zeros([100,32,64,1])
dt=24
dt_virtual=12
count=0
for t in range(0, time, dt_virtual):
if (t==0):
u = model_vir.predict(TRUTH[0,:].reshape([1,32,64,1]))
else:
u=model_vir.predict(u.reshape([1, 32, 64, 1]))
virtual_obs[count,:,:,0]=np.reshape(u,[32,64])
count=count+1
vir_obs_rs=np.reshape(virtual_obs,[np.size(virtual_obs,0), np.size(virtual_obs,1)*np.size(virtual_obs,2)])
n=int(32*64)
P=np.eye(n,n)
Q=0.03*np.eye(n,n)
R=0.0001
u_ensemble=np.zeros([32*64,2*32*64])
pred=np.zeros([time,32,64,1])
dt=24
dt_virtual=12
count=0
for t in range(0, time, dt):
for kk in range(0,dt):
if (kk==0):
u=Z_rs[t+kk,:].reshape([1, 32, 64, 1 ])
u=model.predict(u.reshape([1,32,64,1]))
elif (kk == dt_virtual-1):
y = u
y, P = ENKF(y, 2048, P, Q, R, vir_obs_rs[int((t+dt_virtual)/dt_virtual),:], model,u_ensemble)
u = y
else :
u=model.predict(u.reshape([1, 32, 64, 1]))
pred[count,:,:,0]=np.reshape(u,[32,64])
count=count+1
x = u
x, P = ENKF(x, 2048, P, Q, R, Z_rs[t+dt,:], model,u_ensemble)
print('output shape of ENKF', np.shape(x))
# pred[count,:,:,0]=np.reshape(x,[32,64])
# count=count+1
sio.savemat('Version2_DA_every24HR_lead1200_everytime_noise_' + str(noise)+ '_2DA.mat',dict([('prediction',pred),('truth',np.reshape(TRUTH,[np.size(Z_rs,0),32,64,1])),('noisy_obs',np.reshape(Z_rs,[np.size(Z_rs,0),32,64,1]))]))
#sio.savemat('normalize_mean_std.mat',dict([('MEAN',M),('STD',sdev)]))
print('Done writing file')