/
SSDMD_GA762_2022_single.m
227 lines (202 loc) · 6.98 KB
/
SSDMD_GA762_2022_single.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
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%
% Scale separated DMD on the GA762 2022 dataset
%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
clear; close all; clc;
% Load sounder data
station = 'GA762';
file_date = '2022-06-01(015)';
data = load(strcat(station, '_Digisonde_', file_date));
h_low = 1;
h_high = 350;
ne_full = data.ne(h_low:h_high, :);
heights = data.heights(h_low:h_high);
full_times = data.times;
% Load IRI
iri = load(strcat(station, '_IRI_', file_date));
iri_hmin = find(iri.alt_km==heights(1));
iri_hmax = find(iri.alt_km==heights(end));
ne_iri = iri.ne(iri_hmin:iri_hmax, :);
fof2_iri = iri.foF2;
hmf2_iri = iri.hmF2;
% Load Didbase profile characteristics
chars = load(strcat(station, '_FastChars_', file_date));
fof2_data = chars.fof2;
hmf2_data = chars.hmf2;
% Some parameters for the SSDMD model
day = 192;
dt = 1./day;
wave_type = 'coif4';
dmd_tol = -6.0;
corr_tol = -1.95;
nd_train = 10;
nd_test = 2;
nd = nd_train + nd_test;
wave_levels = floor(log2(nd_train*day + 1));
start_ix = 1;
stop_ix = nd*day;
% Run model
fof2_data = fof2_data(start_ix:stop_ix);
fof2_iri = fof2_iri(start_ix:stop_ix);
hmf2_data = hmf2_data(start_ix:stop_ix);
hmf2_iri = hmf2_iri(start_ix:stop_ix);
ne_iri = ne_iri(:, start_ix:stop_ix);
ne = ne_full(:, start_ix:stop_ix);
full_times = full_times(start_ix:stop_ix, :);
[num_rows, num_cols] = size(ne);
% Split sounder train/test data
train_stop_ix = nd_train*day;
train_data = ne(:, 1:train_stop_ix);
test_start_ix = train_stop_ix + 1;
test_stop_ix = train_stop_ix + nd_test*day;
% Fit SSDMD model and make forecast over test data range
model = ssdmd(train_data, wave_levels, wave_type, dmd_tol, ...
corr_tol, day, dt, num_cols, heights);
% Save forecasted parameters for this period
fof2_ssdmd_test = model.fof2(test_start_ix:test_stop_ix);
hmf2_ssdmd_test = model.hmf2(test_start_ix:test_stop_ix);
% Set the test data for this period
fof2_data_test = fof2_data(test_start_ix:test_stop_ix);
hmf2_data_test = hmf2_data(test_start_ix:test_stop_ix);
% Set the IRI data for this period
fof2_iri_test = fof2_iri(test_start_ix:test_stop_ix);
hmf2_iri_test = hmf2_iri(test_start_ix:test_stop_ix);
% Compute MAEs
fof2_ssdmd_mae = mean(abs(fof2_data_test - fof2_ssdmd_test), 'omitnan');
hmf2_ssdmd_mae = mean(abs(hmf2_data_test - hmf2_ssdmd_test), 'omitnan');
fof2_iri_mae = mean(abs(fof2_data_test - fof2_iri_test), 'omitnan');
hmf2_iri_mae = mean(abs(hmf2_data_test - hmf2_iri_test), 'omitnan');
%%
% SSDMD
% Stats
rmse_fof2_ssdmd = sqrt(mean((fof2_data_test - fof2_ssdmd_test).^2, 'omitnan'));
rmse_hmf2_ssdmd = sqrt(mean((hmf2_data_test - hmf2_ssdmd_test).^2, 'omitnan'));
mae_fof2_ssdmd = mean(abs(fof2_data_test - fof2_ssdmd_test), 'omitnan');
mae_hmf2_ssdmd = mean(abs(hmf2_data_test - hmf2_ssdmd_test), 'omitnan');
mape_fof2_ssdmd = mean(abs((fof2_data_test - fof2_ssdmd_test)./fof2_data_test), 'omitnan')*100;
mape_hmf2_ssdmd = mean(abs((hmf2_data_test - hmf2_ssdmd_test)./hmf2_data_test), 'omitnan')*100;
fprintf('\n\tSSDMD stats\n');
fprintf('RMSE fof2: %f\n', rmse_fof2_ssdmd);
fprintf('MAE fof2: %f\n', mae_fof2_ssdmd);
fprintf('MAPE fof2: %f\n', mape_fof2_ssdmd);
fprintf('RMSE hmf2: %f\n', rmse_hmf2_ssdmd);
fprintf('MAE hmf2: %f\n', mae_hmf2_ssdmd);
fprintf('MAPE hmf2: %f\n', mape_hmf2_ssdmd);
% IRI
% Stats
rmse_fof2_iri = sqrt(mean((fof2_data_test - fof2_iri_test).^2, 'omitnan'));
rmse_hmf2_iri = sqrt(mean((hmf2_data_test - hmf2_iri_test).^2, 'omitnan'));
mae_fof2_iri = mean(abs(fof2_data_test - fof2_iri_test), 'omitnan');
mae_hmf2_iri = mean(abs(hmf2_data_test - hmf2_iri_test), 'omitnan');
mape_fof2_iri = mean(abs((fof2_data_test - fof2_iri_test)./fof2_data_test), 'omitnan')*100;
mape_hmf2_iri = mean(abs((hmf2_data_test - hmf2_iri_test)./hmf2_data_test), 'omitnan')*100;
fprintf('\n\tIRI stats\n');
fprintf('RMSE fof2: %f\n', rmse_fof2_iri);
fprintf('MAE fof2: %f\n', mae_fof2_iri);
fprintf('MAPE fof2: %f\n', mape_fof2_iri);
fprintf('RMSE hmf2: %f\n', rmse_hmf2_iri);
fprintf('MAE hmf2: %f\n', mae_hmf2_iri);
fprintf('MAPE hmf2: %f\n', mape_hmf2_iri);
%%
save('Run_GA762_2022')
%% Plot forecast for station
saveFigs = false;
dpi = 150;
ms = 20;
lw = 6;
fs = 28;
% fof2
figure
tl = tiledlayout(2, 1, 'TileSpacing','tight', 'Padding', 'tight');
nexttile;
hold on
plot(fof2_data_test, 'r.', 'MarkerSize', ms)
plot(fof2_ssdmd_test, 'b-', 'LineWidth', lw)
plot(fof2_iri_test, 'k-', 'LineWidth', lw)
grid on;
hold off
title('Station GA762 forecast', 'FontSize', fs)
labelssdmd = sprintf('SSDMD (MAE: %.2f)', fof2_ssdmd_mae);
labeliri = sprintf('IRI (MAE: %.2f)', fof2_iri_mae);
legend('Measured', labelssdmd, labeliri, 'Location','northwest')
ylabel('foF2 (MHz)', 'FontWeight', 'bold')
xticks([0.5*day:day:nd_test*day-0.5*day])
xticklabels([datestr(makeDate(iri.times(test_start_ix:day:end)), 'yyyy/mm/dd')])
h = gca;
h.FontSize = fs;
% hmf2
nexttile;
hold on
plot(hmf2_data_test, 'r.', 'MarkerSize', ms)
plot(hmf2_ssdmd_test, 'b-', 'LineWidth', lw)
plot(hmf2_iri_test, 'k-', 'LineWidth', lw)
grid on;
hold off
labelssdmd = sprintf('SSDMD (MAE: %.2f)', hmf2_ssdmd_mae);
labeliri = sprintf('IRI (MAE: %.2f)', hmf2_iri_mae);
legend('Measured', labelssdmd, labeliri, 'Location','northwest')
ylabel('hmF2 (km)', 'FontWeight', 'bold')
xticks([0.5*day:day:nd_test*day-0.5*day])
xticklabels([datestr(makeDate(iri.times(test_start_ix:day:end)), 'yyyy/mm/dd')])
h = gca;
h.FontSize = fs;
%% Total reconstructions with hmf2 param plotted on top
fs = 36;
% Shift hmf2 so the indices line up with data in plot
hmf2_ = hmf2_data - double(heights(1));
tiledlayout(2, 1, 'TileSpacing','Compact', 'Padding', 'Compact');
nexttile;
s1 = surf(ne);
hold on;
shading flat;
view([0, 90]);
z_max = max(max(ne)+1)*ones(num_cols);
x_line = linspace(1, num_cols, num_cols);
hl = line(x_line, hmf2_, z_max, 'Color', 'k', 'LineWidth', 2);
c = colorbar;
colormap('jet')
c.Label.String = 'Plasma Frequency (MHz)';
caxis([0 6]);
xl = xline(nd_train*day, 'm--', 'linewidth', 4);
title('Measurement', 'FontSize', fs);
h = set(gca,'FontSize', fs);
set(h,'Interpreter','LaTeX')
xticks([1:day:num_cols])
xticklabels([])
xlim([0, num_cols])
ylim([0, num_rows])
xlabel([])
yticks([1:100:num_rows])
yticklabels([heights(1:100:end)])
ylabel('Height (km)', 'FontSize', fs)
legend([hl], 'hmF2')
hold off
% Shift hmf2 so the indices line up with data in plot
hmf2_ = model.hmf2 - double(heights(1));
nexttile;
s2 = surf(model.full_recon);
hold on;
shading flat;
view([0, 90]);
z_max = max(max(ne)+1)*ones(num_cols);
x_line = linspace(1, num_cols, num_cols);
hl = line(x_line, hmf2_, z_max, 'Color', 'k', 'LineWidth', 2);
c = colorbar;
colormap('jet')
c.Label.String = 'Plasma Frequency (MHz)';
caxis([0 6]);
xl = xline(nd_train*day, 'm--', 'linewidth', 4);
title('SSDMD', 'FontSize', fs);
h = set(gca,'FontSize', fs);
set(h,'Interpreter','LaTeX')
xticks([1:day:num_cols])
xticklabels([0:nd])
xlim([0, num_cols])
ylim([0, num_rows])
xlabel('Time (days)', 'FontSize', fs)
yticks([1:100:num_rows])
yticklabels([heights(1:100:end)])
ylabel('Height (km)', 'FontSize', fs)
legend([hl], 'hmF2')
hold off;