-
Notifications
You must be signed in to change notification settings - Fork 16
/
a6_fitbessel_LRT.m
461 lines (390 loc) · 16.3 KB
/
a6_fitbessel_LRT.m
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
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
% Extract phase velocity dispersion between station pairs by fitting J0 bessel
% function to real(ccf)
% Uses cross spectral fitting technique of Menke & Jin (2015) BSSA
% DOI:10.1785/0120140245
%
% Define own starting phase velocity dispersion c manually or using
% functions/calc_Rayleigh_disp for a simple layered model (does not work for
% models with a water column).
%
%
% https://github.com/jbrussell
clear
close all;
global tN
global waxis
global twloc
global weight
setup_parameters;
%======================= PARAMETERS =======================%
% RAYLEIGH FUND MODE
comp = {'ZZ'}; %'RR'; 'ZZ'; 'TT'
windir = 'window3hr';
xspdir = 'ZZ_0S_LRT'; % output directory of phase velocities
N_wl = 1; % Number of wavelengths required
Ninterp = 40; % [] or Number of points to interpolate to;
% Use picks from Linear Radon Transform? (./mat-LRTdisp/)
is_LRT_picks = 1; % Use picks from Radon Transform to determine starting dispersion model and frequencies
LRT_method = 'CGG_weight';
mode_br = 0; % desired mode branch [0=fund.]
frange_LRT = [1/40 1/3]; % Frequency range of LRT panel for reading in picks
frange_fit = [1/25 1/3]; % Frequency range to fit over! Can be more restrictive than where picks were made
damp = [1; 1; 1]; % [fit, smoothness, slope]
is_normbessel = 0; % normalize bessel function by analytic envelope?
if ~is_LRT_picks
frange_fit = [1/40 1/10]; % frequency range over which to fit bessel function
% xlims = [1/70 1/9];
Npers = 18; % Number of periods
t_vec_all = 1./flip(linspace(frange_fit(1) , frange_fit(2) ,Npers)); % periods at which to extract phase velocity
end
xlims = [min(frange_fit)*0.9 max(frange_fit)*1.1]; % limits for plotting
is_resume = 0; % Resume from last processed file (1) or overwrite (0)
iswin = 1; % Use the time-domain windowed ccfs?
npts_smooth = 1; % Smoothing of bessel function: 1 = no smoothing
isoutput = 1; % Save *.mat file with results?
nearstadist = 0;
IsFigure = 1;
isfigure2 = 0;
isfigure_snr = 1;
%% Make initial guess at phase velocity dispersion model
if is_LRT_picks
% Read picks from Linear Radon Transform (see ./mat-LRTdisp/)
ccfstr = strsplit(parameters.ccfpath,'/');
ccfstr = ccfstr{end-1};
in_LRTpicks = [parameters.path_LRT_picks,ccfstr,'/',windir,'/',num2str(1/frange_LRT(2)),'_',num2str(1/frange_LRT(1)),'s/LRTpicks_',LRT_method,'_',comp{1},'.mat'];
temp = load(in_LRTpicks);
picks_LRT = temp.picks_LRT;
imode = mode_br+1;
% Interpolate to number of desired points
if ~isempty(Ninterp)
t_vec_int = 1./(linspace( 1./min(picks_LRT(imode).per), 1./max(picks_LRT(imode).per), Ninterp ));
picks_LRT(imode).phv = interp1( picks_LRT(imode).per, picks_LRT(imode).phv, t_vec_int );
picks_LRT(imode).phv_std = interp1( picks_LRT(imode).per, picks_LRT(imode).phv_std, t_vec_int );
picks_LRT(imode).per = t_vec_int;
end
% Avoid overlapping periods from other mode_br branches
if mode_br == 0
if imode < length(picks_LRT)
I_good = picks_LRT(imode).per > picks_LRT(imode+1).per(end);
else
I_good = true(size(picks_LRT(imode).per));
end
elseif mode_br > 0
if imode-1 ~= 0 && imode < length(picks_LRT)
I_good = (picks_LRT(imode).per < picks_LRT(imode-1).per(1)) & (picks_LRT(imode).per > picks_LRT(imode+1).per(end));
end
if imode-1 ~= 0
I_good = picks_LRT(imode).per < picks_LRT(imode-1).per(1);
end
if imode < length(picks_LRT)
I_good = picks_LRT(imode).per > picks_LRT(imode+1).per(end);
end
end
% Fit only periods within frange_fit
I_fit = picks_LRT(imode).per>=1/frange_fit(2) & picks_LRT(imode).per<=1/frange_fit(1);
I_good = logical(I_fit .* I_good);
c_all = picks_LRT(imode).phv(I_good);
t_vec_all = picks_LRT(imode).per(I_good);
c_all_std = picks_LRT(imode).phv_std(I_good);
c_start = c_all;
else
% Read from MINEOS .q file (https://github.com/jbrussell/MINEOS_synthetics)
qfile = ['./qfiles/Nomelt_taper_eta_crust_INVpconstr_xi1.06_GRL19_ORCAiso_INV.s0to200.q'];
if exist('c','var') == 0 % check if phase velocities exist, if not read them in
[~,~,c_all] = readMINEOS_qfile2(qfile,t_vec_all,mode_br);
end
c_start = c_all;
c_all_std = zeros(size(c_all));
end
%%
%==========================================================%
% LIMITS
if comp{1}(1) == 'R'
ylims = [3.2 4.5];
elseif comp{1}(1) == 'Z' || comp{1}(1) == 'P'
% ylims = [1.5 5.0];
% ylims = [2 4.5];
ylims = [1.5 4.5];
elseif comp{1}(1) == 'T'
ylims = [3.5 4.8];
end
% LOAD DATA TO SEE HOW MANY POINTS
%%% --- Load in the ccf --- %%%
%ccf_path = ['./ccf/',windir,'/fullStack/ccf',comp{1},'/'];
ccf_path = [parameters.ccfpath,windir,'/fullStack/ccf',comp{1},'/'];
stalist = parameters.stalist;
sta1=char(stalist(1,:));
sta2=char(stalist(2,:));
sta1dir=[ccf_path,sta1]; % dir to have all cross terms about this central station
filename = sprintf('%s/%s_%s_f.mat',sta1dir,sta1,sta2);
if ~exist(filename,'file')
disp(['not exist ',filename])
% continue;
end
data1 = load(filename);
npts = length(data1.coh_sum_win);
% input path
%ccf_path = ['./ccf/',windir,'/fullStack/ccf',comp{1},'/'];
ccf_path = [parameters.ccfpath,windir,'/fullStack/ccf',comp{1},'/'];
% output path
XSP_path = ['./Xsp/',windir,'/fullStack/Xsp',comp{1},'/',num2str(1/frange_fit(2)),'_',num2str(1/frange_fit(1)),'s_',num2str(N_wl),'wl_',xspdir,'/'];
if ~exist(XSP_path)
if ~exist('./Xsp/')
mkdir('./Xsp/');
end
if ~exist(['./Xsp/',windir,'/'])
mkdir(['./Xsp/',windir,'/']);
end
if ~exist(['./Xsp/',windir,'/fullStack/'])
mkdir(['./Xsp/',windir,'/fullStack/']);
end
if ~exist(['./Xsp/',windir,'/fullStack/Xsp',comp{1},'/'])
mkdir(['./Xsp/',windir,'/fullStack/Xsp',comp{1},'/']);
end
mkdir(XSP_path)
end
% figure output path
if iswin
XSP_fig_path = ['./figs/',windir,'/fullStack/Xsp/',num2str(1/frange_fit(2)),'_',num2str(1/frange_fit(1)),'s_',num2str(N_wl),'wl_',xspdir,'/TEI19/'];
else
XSP_fig_path = ['./figs/',windir,'/fullStack/Xsp/',num2str(1/frange_fit(2)),'_',num2str(1/frange_fit(1)),'s_',num2str(N_wl),'wl_',xspdir,'/TEI19_nowin/'];
end
if ~exist(XSP_fig_path)
mkdir(XSP_fig_path);
end
warning off; %#ok<WNOFF>
stalist = parameters.stalist;
nsta=parameters.nsta; % number of target stations to calculate for
%%% --- Loop through station 1 --- %%%
for ista1=1:nsta
sta1=char(stalist(ista1,:));
sta1dir=[ccf_path,sta1]; % dir to have all cross terms about this central station
%%% --- Loop through station 2 --- %%%
for ista2 = 1: nsta % length(v_sta)
sta2 = char(stalist(ista2,:));
% if same station, skip
if(strcmp(sta1,sta2))
continue
end
% Check to see if we have already done this
if is_resume && exist([XSP_path,sta1,'_',sta2,'_xsp.mat'])
disp('Already fit this one!')
continue
end
clear data1 xcorf1 xsp1 filename
%%% --- Load in the ccf --- %%%
filename = sprintf('%s/%s_%s_f.mat',sta1dir,sta1,sta2);
if ~exist(filename,'file')
disp(['not exist ',filename])
continue;
end
data1 = load(filename);
r1 = distance(data1.stapairsinfo.lats(1),data1.stapairsinfo.lons(1),data1.stapairsinfo.lats(2),data1.stapairsinfo.lons(2),referenceEllipsoid('GRS80'))/1000;
dt = data1.stapairsinfo.dt;
groupv_max = data1.max_grv;
groupv_min = data1.min_grv;
if r1 < nearstadist
continue;
end
% Index wavelength criterion
I_wl = r1 ./ (t_vec_all .* c_all) > N_wl;
if sum(I_wl) <= 1
I_wl(1) = 1;
I_wl(2) = 1;
end
c = c_all(I_wl);
t_vec = t_vec_all(I_wl);
c_std = c_all_std(I_wl);
tN = length(t_vec);
wholesec = npts;
wvec1 = (2*pi)./t_vec;
wvec1 = wvec1';
% Get your axis correct
twloc=1./t_vec;
twloc = twloc*2*pi;
% waxis = (frange_fit(1):1/wholesec:frange_fit(2))*2*pi;
waxis = (1/max(t_vec):1/wholesec:1/min(t_vec))*2*pi;
%%% - Get the normalized ccf - %%%
if iswin
xcorf1 = data1.coh_sum_win./data1.coh_num;
else
xcorf1 = data1.coh_sum./data1.coh_num;
end
dumnan = find(isnan(xcorf1)==1);
if length(dumnan) > 10
disp([sta1,' and ',sta2,'is NaN! Moving on']);
continue
end
%N = 10000;
N = length(xcorf1);
if length(xcorf1) < N
disp('Dataset is too short! Moving on')
continue
end
xcorf = xcorf1;
xcorf1 = real(xcorf1(1:N));
xcorf1(1) = 0;
if isfigure2
figure(1)
T = length(xcorf1);
temp_faxis = [0:1/T:1/dt/2,-1/dt/2+1/T:1/T:-1/T];
ind = find(temp_faxis>0);
subplot(2,1,1)
%plot(temp_faxis(ind),smooth(real(xcorf1(ind)),100));
plot(flip(temp_faxis(ind),smooth(real(xcorf1(ind)),50)));
xlim([frange_fit(1) frange_fit(2)])
hold on
subplot(2,1,2)
%plot(temp_faxis(ind),smooth(real(data1.coh_sum(ind)/data1.coh_num),100),'-r')
plot(temp_faxis(ind),smooth(real(data1.coh_sum(ind)/data1.coh_num),50),'-r');
xlim([frange_fit(1) frange_fit(2)])
end
%%% - Convert xcorf into spherical frequency - %%%
faxis = [0:(N-mod(N-1,2))/2 , -(N-mod(N,2))/2:-1]/dt/N;
xsp1 = interp1(faxis*2*pi,xcorf1,waxis);
%xsp1 = smooth(xsp1,50);
xsp1 = smooth(xsp1,npts_smooth);
tw1 = ones(1,tN)*r1./c;
%%% - Invert for the bessel function 2x - %%%
options = optimoptions(@lsqnonlin,'TolFun',1e-12,'MaxIter',1500,'MaxFunEvals',1500);
weight = 1./waxis;
tw2 = lsqnonlin(@(x) besselerr(x,[xsp1],damp,is_normbessel),[tw1],[tw1]*0.8,[tw1]*1.2,options);
% tw2 = lsqnonlin(@(x) besselerr(x,[xsp1]),[tw1],[],[],options);
weight(:) = 1;
[tw,~,res,~,~,~,J] = lsqnonlin(@(x) besselerr(x,[xsp1],damp,is_normbessel),[tw2],[tw2]*0.8,[tw2]*1.2,options);
% tw = lsqnonlin(@(x) besselerr(x,[xsp1]),[tw2],[tw2]*0.8,[tw2]*1.2,options);
% tw = lsqnonlin(@(x) besselerr(x,[xsp1]),[tw2],[],[],options);
% ESTIMATE ERROR BARS ON MODEL! :: JBR - 2/2020
% Calculate data variance from residual following Menke QMDA book eq. (4.31)
% s_d^2 = E / (N-M)
sigma_d2 = res'*res / (length(res)-length(tw));
Cov_m = inv(J'*J)*sigma_d2;
sigma_m_tw = diag(Cov_m).^(1/2);
% Propogate model error to phase velocity
% c = r1./tw; therefore dc = |r*t^(-2) * dt|
sigma_m_c = abs(r1.*tw'.^(-2).*sigma_m_tw);
%%% - Set up the variable structure - %%%
xspinfo.sta1 = sta1;
xspinfo.sta2 = sta2;
xspinfo.lat1 = data1.stapairsinfo.lats(1);
xspinfo.lon1 = data1.stapairsinfo.lons(1);
xspinfo.lat2 = data1.stapairsinfo.lats(2);
xspinfo.lon2 = data1.stapairsinfo.lons(2);
xspinfo.r = r1;
xspinfo.tw = tw;
xspinfo.xsp = xsp1;
xspinfo.xsp_norm = xsp1./abs(hilbert(xsp1));
xspinfo.coherenum = data1.coh_num;
err = besselerr(tw,xsp1,damp,is_normbessel);
err = err(1:length(waxis));
if is_normbessel
xspinfo.sumerr = sum(err.^2)./sum((xspinfo.xsp_norm./weight(:)).^2);
else
xspinfo.sumerr = sum(err.^2)./sum((xsp1./weight(:)).^2);
end
xspinfo.err = err./weight(:);
xspinfo.tw1 = tw1;
xspinfo.twloc = twloc;
xspinfo.c = r1./tw;
xspinfo.c_std = sigma_m_c;
xspinfo.per = 1./(twloc/2/pi);
xspinfo.c_start = c_start;
xspinfo.c_std_start = c_all_std;
xspinfo.per_start = t_vec_all;
xspinfo.isgood_wl = I_wl;
data = r1./tw;
%% %%% Calculate SNR %%%
xcorf1 = data1.coh_sum./data1.coh_num;
xcorf1_filtered = tukey_filt( xcorf1,[min(t_vec) max(t_vec)],dt,0.25 );
[snr, signal_ind] = calc_SNR(xcorf1_filtered,groupv_min,groupv_max,r1,dt,isfigure_snr);
%%
xspinfo.filename = filename;
xspinfo.snr = snr;
% Calculate the predicted bessel function from the initial model
disp([filename,' fitted'])
if IsFigure
if 0 % plot initial bessel
tw_init = interp1(twloc,tw1(1:tN),waxis,'linear');
x_init = waxis.*tw_init;
A = 1;
binit = besselj(0,x_init)*A;
binit = binit./mean(abs(binit)).*mean([abs(xsp1)]);
plot(waxis/2/pi,binit,'-k','linewidth',2); hold on;
end
f3 = figure(3); clf; hold on;
set(gcf,'color','w','Position',[289 1 517 704]);
ax1 = subplot(3,1,1);
plot_SNR(xcorf1_filtered,groupv_min,groupv_max,r1,dt,ax1);
set(ax1,'box','off');
% REAL PART (J0)
subplot(3,1,2); box off;
tww = interp1(twloc,tw(1:tN),waxis,'linear');
x = waxis.*tww;
A = 1;
b = besselj(0,x)*A;
b = b./mean(abs(b)).*mean([abs(xsp1)]);
T = length(data1.coh_sum);
faxis = [0:1/T:1/dt/2,-1/dt/2+1/T:1/T:-1/T];
ind = find(faxis>0);
plot(faxis(ind),smooth(real(data1.coh_sum_win(ind)/data1.coh_num),npts_smooth),'-k','linewidth',3); hold on;
if ~iswin
plot(waxis/2/pi,xsp1,'-b','linewidth',1);
end
plot(waxis/2/pi,b,'-r','linewidth',2); hold on;
xlim(xlims);
xlims1 = get(gca,'XLim');
ylabel('J_{0}','fontsize',16);
ax1 = get(gca);
dx = 1;
dy = 0.95;
text(gca,diff(ax1.XLim)*dx+ax1.XLim(1),diff(ax1.YLim)*dy+ax1.YLim(1),...
sprintf('%.4f',xspinfo.sumerr),'color',[0 0 0],'fontsize',14);
set(gca,'fontsize',16,'linewidth',1.5);
box off;
title(['Distance : ',num2str(r1),'km'],'fontsize',16);
hold on
subplot(3,1,3);
errorbar(twloc/2/pi,r1./tw1,c_std,'o-','color',[0.5 0.5 0.5],'linewidth',2);hold on;
errorbar(twloc/2/pi,r1./tw,sigma_m_c*2,'o-','color',[1 0 0 ],'linewidth',2);
% errorbar(twloc/2/pi,r1./tw,xspinfo.err,'ro-','linewidth',2);
title([sta1,'-',sta2],'fontsize',16)
xlabel('Frequency (Hz)','fontsize',16);
ylabel('Phase Velocity (km/s)','fontsize',16);
set(gca,'fontsize',16,'linewidth',1.5);
ylim(ylims);
xlim(xlims1);
box off;
% Plot normalized bessel functions
if is_normbessel
figure(4); clf;
b_dat = smooth(real(data1.coh_sum_win(ind)/data1.coh_num),npts_smooth);
plot(faxis(ind),b_dat./abs(hilbert(b_dat)),'k','linewidth',3); hold on;
plot(waxis/2/pi,b./SmoothAnalyticEnv(waxis/2/pi,b),'-r','linewidth',2); hold on;
xlim(xlims);
xlims1 = get(gca,'XLim');
ylabel('J_{0}','fontsize',16);
ylim([-2 2]);
end
if isfigure2
f12 = figure(12);
clf
T = length(data1.coh_sum);
faxis = [0:1/T:1/dt/2,-1/dt/2+1/T:1/T:-1/T];
ind = find(faxis>0);
plot(faxis(ind),smooth(real(data1.coh_sum(ind)/data1.coh_num),npts_smooth),'-r');
xlim([frange_fit(1) frange_fit(2)])
end
psfile = [XSP_fig_path,'Xsp_',comp{1}(1),'_',sta1,'_',sta2,'_J0J1.pdf'];
%print('-dpsc2',psfile);
drawnow
if isoutput
save2pdf(psfile,f3,250);
end
% pause;
end
if isoutput
save(sprintf('%s/%s_%s_xsp.mat',XSP_path,sta1,sta2),'xspinfo','twloc','waxis');
end
% pause;
end %end of station j
end %end of station i