-
Notifications
You must be signed in to change notification settings - Fork 10
/
Copy pathmGlobe_correctionFactor.m
307 lines (293 loc) · 18.1 KB
/
mGlobe_correctionFactor.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
function mGlobe_correctionFactor(Input,dem_file,orography_file,mGlobe_output,local_pressure_tsf,local_pressure_channel)
% MGLOBE_CORRECTIONFACTOR computes the correction factor
% Function serves for the computation of site- and model-dependent
% correction/admittance factor. This factor should be use for the multiplication
% of pressure residuals (in situ - ERA Interim). The corresponding effect must
% be then added to the mGlobe atmospheric effect in order to reduce the
% effect of deficient spatial and temporal resolution.
% This function requires Matlab R2012a or higher including Mapping and
% Statistics toolboxes + 8 GB of RAM. Additionally, this function uses
% several functions of the mGlobe toolbox and therefore should be stored in
% the same folder. To reduce the resolution of the computation grid, modify
% row 56.
%
% Inputs:
% Input ... Coordinates of the gravimeter (WGS84 + height)
% [Latitude (deg), Longitude (deg), Height (m)]
% dem_file ... Matlab structure array file containing digital
% elevation model (dem.lon,dem.lat,dem.height).
% This file can by generated using the mGlobe
% build-in conversion tool. Loading of a DEM with
% insufficient extension will result in NaNs or
% correction factor equal to zero. Maximal
% integration area = 0.1 deg (spherical distance)
% The difference between the Input Height and
% interpolated DEM height is used to shift the
% whole DEM "towards" the gravimeter. If you do
% not want to shift the DEM (e.g. underground
% gravimeter) set Input(3) = NaN.
% orography_file ... NetCDF (ERA) or HDF (MERRA) file containing the
% orography. Same as for the computation of
% mGlobe atmospheric effect.
% mGlobe_output ... txt mGlobe output file (mGlobe-Atmo).
% local_pressure_tsf. TSF file containing local (in situ) pressure
% variations in hPa.
% local_pressure_channel TSF channel containing the pressure variations
% in hPa (channel not the column).
%
% Outputs:
% The result will be printed to the command line. Additional plots will
% be used to visualize the results.
%
% Example:
% Input = [47.9283 15.8598 1044];
% dem_file = 'CO_DEM_upTo_01deg_WGS84.mat';
% orography_file = '\ERAinterim\Invariant\ERA_INVARIANT_ALL.nc';
% mGlobe_output = 'F:\mikolaj\Documents\MYpapers\mGlobe\Results\ATMO\mGlobE\CO\CO_MGLOBE_ATMO_ERA_012010_122013_6h_dem0.txt';
% local_pressure_tsf = 'CO_SG_g_and_p.tsf';
% local_pressure_channel = 2;
% mGlobe_correctionFactor(Input,dem_file,orography_file,mGlobe_output,local_pressure_tsf,local_pressure_channel);
%
% M. Mikolaj, 18.11.2014
%% Main settings
psi = 0.10; % fixed integration radius
resol = 0.0002; % fixed spatial resolution
%% Load input DEM and mGlobe output
fprintf('Reading DEM...\n');
dem = importdata(dem_file);
dem.height = dem.height;
Hdem = interp2(dem.lon,dem.lat,dem.height,Input(2),Input(1)); % height of the compuation point with respect to orography
if ~isnan(Input(3))
dem.height = dem.height + (Input(3)-Hdem); % shift the DEM towards the actual height.
end
meteo = load(mGlobe_output);
time = meteo(:,1); % time
p_era = meteo(:,9); % pressure in Pa
t = meteo(:,10); % temperature in K
q = meteo(:,11); % specific humidity in kg/kg
Input(4) = mGlobe_elip2sphere(Input(2)*pi/180,Input(1)*pi/180,Hdem)*180/pi; % spherical latitude (deg)
%% Load local pressure data
fprintf('Reading TSF...\n');
fid = fopen(local_pressure_tsf);
tline = fgetl(fid);header_lines = 0;
if ~isempty(tline)
while isnan(str2double(tline(1))) % determine number of header lines
tline = fgetl(fid);
try
if strcmp(char(tline(2:9)),'UNDETVAL') % try to find flagged values
flagg = tline(11:end);
flagg = str2double(flagg);
end
end
if isempty(tline)
tline = 'mGlobe_RULEZ';
end
header_lines = header_lines + 1; % count header lines
end
fclose(fid); % close the file to open it with known number of header lines
grav = dlmread(local_pressure_tsf,'',header_lines,0); % warning no footer info are allowed
grav(grav(:,local_pressure_channel+6)==flagg,local_pressure_channel+6) = NaN; % remove flagged values
p_obs = interp1(datenum(grav(:,1:6)),grav(:,local_pressure_channel+6)*100,time); % interpolate the local pressure+convert to Pa
%% Load Orography
fprintf('Reading Orography...\n');
switch orography_file(end-1:end)
case 'nc'
ncid_ref = netcdf.open(orography_file,'NC_NOWRITE'); % open given netCDF files
[ndims,nvars] = netcdf.inq(ncid_ref); % get variable names
for nc = 1:nvars
varname = netcdf.inqVar(ncid_ref,nc-1);
switch varname
% MERRA
case 'XDim'
orography.lon = netcdf.getVar(ncid_ref,nc-1,'double'); % get longitude
case 'YDim'
orography.lat = netcdf.getVar(ncid_ref,nc-1,'double'); % get latitude
case 'PHIS'
orography.data = netcdf.getVar(ncid_ref,nc-1,'double'); % get orography
try
scale_factor = netcdf.getAtt(ncid_ref,nc-1,'scale_factor'); % check if scaling factor does exist
catch
scale_factor = 1;
end
try
add_offset = netcdf.getAtt(ncid_ref,nc-1,'add_offset'); % check if offset does exist
catch
add_offset = 0;
end
orography.data = orography.data'*scale_factor + add_offset; % scale and transpose the orography matrix
orography.data = double(orography.data)./9.80665; % transform to meters using CONSTANT gravity (no latitude dependency!)
[orography.lon,orography.lat] = meshgrid(double(orography.lon),double(orography.lat)); % vectors to matrices
% ERA
case 'longitude'
orography.lon = netcdf.getVar(ncid_ref,nc-1,'double'); % get longitude
case 'latitude'
orography.lat = netcdf.getVar(ncid_ref,nc-1,'double'); % get latitude
case 'z'
orography.data = netcdf.getVar(ncid_ref,nc-1,'double'); % get orography
try
scale_factor = netcdf.getAtt(ncid_ref,nc-1,'scale_factor'); % check if scaling factor does exist
catch
scale_factor = 1;
end
try
add_offset = netcdf.getAtt(ncid_ref,nc-1,'add_offset'); % check if offset does exist
catch
add_offset = 0;
end
orography.data = orography.data'*scale_factor + add_offset; % scale and transpose the orography matrix
orography.data = orography.data./9.80665; % transform to meters using CONSTANT gravity (no latitude dependency!)
[orography.lon,orography.lat] = meshgrid(orography.lon,orography.lat); % create lon/lat matrices
clear ndims nvars nc add_offset scale_factor % remove used variables
orography.lon(orography.lon>=180) = orography.lon(orography.lon>=180) - 360; % transform coordinates to (-180,180) system
ri = find(abs(diff(orography.lon(1,:)))==max(abs(diff(orography.lon(1,:)))));
orography.lon = horzcat(orography.lon(:,ri+1:end),orography.lon(:,1:ri)); % Connect matrices to remove discontinuity
orography.lat = horzcat(orography.lat(:,ri+1:end),orography.lat(:,1:ri));
orography.data = horzcat(orography.data(:,ri+1:end,:),orography.data(:,1:ri,:));
end
end
netcdf.close(ncid_ref); % geopotential to height
case 'df' % MERRA only
orography.data = hdfread(orography_file,'PHIS'); % read hdf file containing MERRA orography
orography.data(orography.data>1000000) = NaN; % set flagged values to NaN (no flagged values are expected though)
orography.data = double(orography.data)./9.80665; % convert to m
orography.lon = hdfread(orography_file,'XDim:EOSGRID'); % longitude vector
orography.lat = hdfread(orography_file,'YDim:EOSGRID'); % latitude vector
[orography.lon,orography.lat] = meshgrid(double(orography.lon),double(orography.lat)); % vectors to matrices
end
%% Create grid
fprintf('Preparing computation...\n');
psi_lon = 0.05;delta_lon = 0.05;
while psi_lon<psi
delta_lon = delta_lon + 0.02;
psi_lon = acos(sin(Input(4)*pi/180).*sin(Input(4)*pi/180) + cos(Input(4)*pi/180).*cos(Input(4)*pi/180).*cos((delta_lon)*pi/180))*180/pi;
end
psi_lat = 0.05;delta_lat = 0.05;
while psi_lat<psi
delta_lat = delta_lat + 0.02;
psi_lat = acos(sin(Input(4)*pi/180).*sin((Input(4)-delta_lat)*pi/180) + cos(Input(4)*pi/180).*cos((Input(4)-delta_lat)*pi/180).*cos(0))*180/pi;
end
[comp.lon,comp.lat] = meshgrid(Input(2)-delta_lon-resol/2:resol:Input(2)+delta_lon+resol/2,... % computation grid for spherical coordinated
Input(4)-delta_lat-resol/2:resol:Input(4)+delta_lat+resol/2);
% Transform to sphere
a = 6378137; % ellipsoidal major axis (m)
b = 6356752.314245;
e = sqrt((a^2-b^2)/a^2);e2c = (a^2-b^2)/b^2;
R = sqrt(1+2/3*e^2+3/5*e^4+4/7*e^6+5/9*e^8+6/11*e^10+7/13*e^12)*b; % Radius of the replacement sphere -> equal surface
X = (R).*cos(comp.lat*pi/180).*cos(comp.lon*pi/180); % approximated solution (zero heights)
Y = (R).*cos(comp.lat*pi/180).*sin(comp.lon*pi/180);
Z = (R).*sin(comp.lat*pi/180);
p = sqrt(Y.^2+X.^2);
tet = atan((Z.*a)./(p.*b));
comp.lat_ell = atan((Z+e2c*b*(sin(tet)).^3)./(p-e^2*a*(cos(tet)).^3));
clear tet p X Y Z e a b e2c % remove used variables
comp.lat_ell = comp.lat_ell*180/pi;
comp.dem = interp2(dem.lon,dem.lat,dem.height,comp.lon,comp.lat_ell); % get topography heights
comp.oro = interp2(orography.lon,orography.lat,orography.data,comp.lon,comp.lat_ell); % get orography heights
comp.psi = acos(sin(comp.lat*pi/180).*sin(Input(4)*pi/180) + cos(comp.lat*pi/180).*cos(Input(4)*pi/180).*cos((Input(2)-comp.lon)*pi/180)); % spherical distance
Hdem = interp2(dem.lon,dem.lat,dem.height,Input(2),Input(1)); % height of the compuation point with respect to orography
Horo = interp2(orography.lon,orography.lat,orography.data,Input(2),Input(1)); % height of the compuation point with respect to orography
height = comp.dem-comp.oro; % tesseroid heights
clear dem orography; % remove used variables
%% Determine the density of each tesseroid
lower_boundary = comp.oro; % create lower boundary variable
lower_boundary(height<0) = comp.dem(height<0); % new lower boundary for tesseroids where orography is above topography
dens = height; dens(:,:) = 1; % create density variable
dens(height<0) = -1; % negative density for tesseroids below the orography
dens(comp.psi*180/pi>psi) = 0; % remove tesseroids outside the local zone
comp.dem = [];
%% Show altitude differences
% figure
% mesh(comp.lon,comp.lat,comp.oro);view(0,90);colorbar;
% title('orography');
%
% figure
% mesh(comp.lon,comp.lat,comp.dem);view(0,90);colorbar;
% title('DEM');
%
% figure
% mesh(comp.lon,comp.lat,lower_boundary);view(0,90);colorbar;
% title('Lower boundary');
%
% figure
% mesh(comp.lon,comp.lat,dens);view(0,90);colorbar;
% title('density matrix');
% comp.psi = []; comp.dem = [];
%
% figure
% mesh(comp.lon,comp.lat,height);view(0,90);colorbar;
% title('DEM-orography');
%% Compute gravitational effect
fprintf('Computing graivty effect (DEM-Orography)...\n');
dg = mGlobe_tesseroid([Input(2),Input(4),Hdem],comp.lon,comp.lat,lower_boundary,abs(height),dens,resol,resol,R,0);
dg = dg*1e+9; % convert to nm/s^2
dg(comp.psi*180/pi>psi) = 0; % set points outside the integration radius to zero
clear dens;
%% Gravitational effect for each time epoch
effect(1:length(time),1) = NaN;
p_rep(1:length(time),1) = NaN;
clc
for i = 1:1:length(time)
fprintf('Computing correction factor... (%3.0f%%)\n',round((i/length(time)*100)));
p_down = p_era(i)*exp((9.81*(comp.oro-lower_boundary))./(287*t(i)*(1+0.608*q(i)))); % pressure for lower boundary
p_up = p_era(i)*exp((9.81*(comp.oro-(lower_boundary+abs(height))))./(287*t(i)*(1+0.608*q(i)))); % pressure for upper boundary
p_rep(i) = p_era(i)*exp((9.81*(Horo-Hdem))./(287*t(i)*(1+0.608*q(i)))); % pressure estimation for gravimeter position
t_down = t(i)+(comp.oro-lower_boundary)*-0.0065; % temperature at the lower boundary
t_up = t(i)+(comp.oro-(lower_boundary+abs(height)))*-0.0065; % temperature at the upper boundary
density_up = p_up./(287.*t_up.*(1-q(i)+q(i)/0.62197)); % density at the upper boundary
density_down = p_down./(287.*t_down.*(1-q(i)+q(i)/0.62197)); % density at the lower boundary
density_up(comp.psi*180/pi>psi) = 0; % set points outside the integration radius to zero
density_down(comp.psi*180/pi>psi) = 0; % set points outside the integration radius to zero
clear p_up p_down
effect(i,1) = sum(sum(dg.*abs((density_up+density_down)/2))); % total gravity effect for the mean density
clc
end
%% Estimate the correction factor
delta_p = (p_obs-p_era)/100;
x = delta_p - mean(delta_p(~isnan(delta_p))); % x values = observed pressure - orography
if Hdem > Horo
y = effect - mean(effect(~isnan(effect))); % y = function of x
else
y = - effect + mean(effect(~isnan(effect))); % y = - effect because of the sign switch related to missing+above lying air
end
r = find(isnan(x)|isnan(y));
if ~isempty(r)
x(r) = [];y(r) = [];
end
poly = polyfit(x,y,1); % compute slope
correction_factor = poly(1);
out_fit = polyval(poly,x); % fitted dp vs dg
%% Plot
figure
subplot(14,1,1:4);
plot(x,y,'k.',x,out_fit,'r');
xlabel('pressure residuals (hPa)');
ylabel('gravity effect (nm s^{-2})');
legend('Results','Fitted');
title('Pressure residuals vs. gravitational effect');
subplot(14,1,7:9)
plot(time,p_obs/100,'k',time,p_era/100,'b',time,p_rep/100,'r');
xlabel('time (matlab)');
ylabel('pressure (hPa)');
legend('Observed','Orography','Estimated');
set(gca,'XTickLabel',[]);
grid on
subplot(14,1,11:14)
plotyy(time,p_rep/100-p_era/100,time,p_rep/100-p_obs/100);
% yl1 = get(h(1),'YLim');
% delta_p0 = p_rep/100-p_obs/100;
% set(h(2),'YLim',[round(mean(delta_p(~isnan(delta_p0)))-(range(yl1))/2),round(mean(delta_p(~isnan(delta_p0)))+(range(yl1))/2)]);
xlabel('time (matlab)');
ylabel('pressure (hPa)');
legend('Estimated-Orography','Estimated-Observed');
grid on
%% Output
if correction_factor < -2 && correction_factor > -4
fprintf('Estimated correction factor = %5.2f nm/s^2/hPa\n',correction_factor);
else
t = meteo(:,4); t = t - mean(t(~isnan(t)));
f = meteo(:,9)/100; f = f - mean(f(~isnan(f)));
c2 = polyfit(f,t,1);
fprintf('The estimated correction factor (=%5.2f) lies outside the expected range, i.e. <-4,-2> nm/s^2/hPa.\nIt is not recommended to use such correction factor. Please check the plotted results.\nThe correct computation requires good performance of the estimated pressure, i.e. the estimated pressure must be similar to the observed air pressure.\nAdditionally, the height interpolated from the loaded DEM must be very similar to the real gravimeter altitude.\nMake sure that the input time series (mGlobe output) are at least 2 years long and sampled every 6 hours.\nThis algorithm fails for areas where the ERA Interim orography is to close to the actual topography (see paper for details).\nValues around 0 nm/s^2/hPa may indicate that the volume above the orography cancels out the effect of the volume below the orography.\nWe recommend to use correction factor %5.2f /s^2/hPa (estimated from least square adjustment of mGlobe total effect and ERA air pressure).\nAlternatively, use site-specific correction factor determined for frequencies higher than 4 cpd.\nOtherwise, the neglected deficient temporal resolution may introduce an error.\n',correction_factor,c2(1));
end
end