Global Index (short | long) | Local contents | Local Index (short | long)
Step 1: Calculate an SST anomaly data set for the 4 by 6 latitude-longitude resolution that Yuan used.
| This script calls | |
|---|---|
% First, define path so my routines work
path('/home/disk/tao/dvimont/matlab/NetCDF',path);
path('/home/disk/tao/dvimont/matlab/Stat_Tools',path);
%%%%%%%%%%%%%%%%%%%%%%%%%%%% Start with SLP %%%%%%%%%%%%%%%%%%%%%
% Load SLP data
cd /home/disk/tao/data/coads/coads1a
filename = 'slp.mean.5097.nc';
varname = 'slp';
time = 'all'; % Equivalently, one may use tim = [1:576], unless
% more months are added.
latlonlim = [0 360 -90 90];
level = 1;
[slp, lat, lon, time] = ...
getnc2(filename, varname, latlonlim, level, 'all');
% Make 4x6 degree averages
[ntim, nlat, nlon] = size(slp);
slp2 = repmat(NaN, [ntim nlat/2 nlon/3]);
for i = 1:nlat/2;
latind = 2*(i-1)+[1:2];
for j = 1:nlon/3;
lonind = 3*(j-1)+[1:3];
slp2(:,i,j) = ...
squeeze(mean2(mean2(shiftdim(slp(:,latind,lonind), 1))));
end
end
lat2 = repmat(NaN, [nlat/3 1]);
for i = 1:nlat/2;
latind = 2*(i-1)+[1:2];
lat2(i) = mean(lat(latind));
end
lon2 = repmat(NaN, [nlon/3 1]);
for j = 1:nlon/3;
lonind = 3*(j-1)+[1:3];
lon2(j) = mean(lon(lonind));
end
% Now, remove climatology, and rename slp2 to slp
[slp, slpclim] = annave(slp2);
% Write slp and slpclim to a netcdf file
% Define (open) NetCDF files
cd /home/disk/tao/data/coads/sstanom4by6
nc_in = netcdf('slpcoadsanom4by6.18541995.nc', 'nowrite');
nc_out = netcdf('slpcoadsanom4by6.19501997.nc', 'clobber');
% Define other dimensions (lat, lon)
nc_out('lat') = length(lat2);
nc_out('lon') = length(lon2);
% Define record dimension (time)
nc_out('time') = 0; d = nc_out('time');
d(:) = length(time);
% Copy variable information (but not data) from nc_in to nc_out
nc_out < nc_in{'lat'};
nc_out{'lat'} < att(nc_in{'lat'});
nc_out < nc_in{'lon'};
nc_out{'lon'} < att(nc_in{'lon'});
nc_out < nc_in{'time'};
nc_out{'time'} < att(nc_in{'time'});
nc_out < nc_in{'data'};
nc_out{'data'} < att(nc_in{'data'});
% Store dimension variable data into nc_out
nc_out{'lat'}(:) = lat2;
nc_out{'lon'}(:) = lon2;
nc_out{'time'}(1:length(time)) = time;
nc_out{'time'}.units(:) = 'days since 1-1-1 00:00:0.0';
% Store slp data into nc_out
mv = nc_in{'data'}.missing_value(:);
ao = nc_in{'data'}.add_offset(:);
sf = nc_in{'data'}.scale_factor(:);
slp2 = (slp-ao)/sf;
slp2(isnan(slp2)) = mv;
nc_out{'data'}(:,:,:) = slp2(:,:,:);
% Add some descriptive attributes
nc_out{'data'}.long_name(:) = ...
'Sea Level Pressure Monthly Mean at Surface';
nc_out{'data'}.dataset = 'COADS 2-degree Enhanced';
nc_out{'data'}.averaging = ...
'2 grid points in latitude, 3 grid points in longitude';
close(nc_out);
close(nc_in);
%%%%%%%%%%%%%%%%%%%%%%%%%%%% Next, SST %%%%%%%%%%%%%%%%%%%%%
% Load SST data
cd /home/disk/tao/data/coads/coads1a
filename = 'sst.mean.5097.nc';
varname = 'sst';
time = 'all'; % Equivalently, one may use tim = [1:576], unless
% more months are added.
latlonlim = [0 360 -90 90];
level = 1;
[sst, lat, lon, time] = ...
getnc2(filename, varname, latlonlim, level, 'all');
% Make 4x6 degree averages
[ntim, nlat, nlon] = size(sst);
sst2 = repmat(NaN, [ntim nlat/2 nlon/3]);
for i = 1:nlat/2;
latind = 2*(i-1)+[1:2];
for j = 1:nlon/3;
lonind = 3*(j-1)+[1:3];
sst2(:,i,j) = ...
squeeze(mean2(mean2(shiftdim(sst(:,latind,lonind), 1))));
end
end
lat2 = repmat(NaN, [nlat/3 1]);
for i = 1:nlat/2;
latind = 2*(i-1)+[1:2];
lat2(i) = mean(lat(latind));
end
lon2 = repmat(NaN, [nlon/3 1]);
for j = 1:nlon/3;
lonind = 3*(j-1)+[1:3];
lon2(j) = mean(lon(lonind));
end
% Now, remove climatology, and rename sst2 to sst
[sst, sstclim] = annave(sst2);
% Write sst and sstclim to a netcdf file
% Define (open) NetCDF files
cd /home/disk/tao/data/coads/sstanom4by6
nc_in = netcdf('sstcoadsanom4by6.18541995.nc', 'nowrite');
nc_out = netcdf('sstcoadsanom4by6.19501997.nc', 'clobber');
% Define other dimensions (lat, lon)
nc_out('lat') = length(lat2);
nc_out('lon') = length(lon2);
% Define record dimension (time)
nc_out('time') = 0; d = nc_out('time');
d(:) = length(time);
% Copy variable information (but not data) from nc_in to nc_out
nc_out < nc_in{'lat'};
nc_out{'lat'} < att(nc_in{'lat'});
nc_out < nc_in{'lon'};
nc_out{'lon'} < att(nc_in{'lon'});
nc_out < nc_in{'time'};
nc_out{'time'} < att(nc_in{'time'});
nc_out < nc_in{'data'};
nc_out{'data'} < att(nc_in{'data'});
% Store dimension variable data into nc_out
nc_out{'lat'}(:) = lat2;
nc_out{'lon'}(:) = lon2;
nc_out{'time'}(1:length(time)) = time;
nc_out{'time'}.units(:) = 'days since 1-1-1 00:00:0.0';
% Store sst data into nc_out
mv = nc_in{'data'}.missing_value(:);
ao = nc_in{'data'}.add_offset(:);
sf = nc_in{'data'}.scale_factor(:);
sst2 = (sst-ao)/sf;
sst2(isnan(sst2)) = mv;
nc_out{'data'}(:,:,:) = sst2(:,:,:);
% Add some descriptive attributes
nc_out{'data'}.long_name(:) = ...
'"Sea Surface Temperature Monthly Mean at Surface';
nc_out{'data'}.dataset = 'COADS 2-degree Enhanced';
nc_out{'data'}.averaging = ...
'2 grid points in latitude, 3 grid points in longitude';
close(nc_out);
close(nc_in);
%%%%%%%%%%%%%%%%%%%%%%%%%%%% Next, UWND %%%%%%%%%%%%%%%%%%%%%
% Load SST data
cd /home/disk/tao/data/coads/coads1a
filename = 'uwnd.mean.5097.nc';
varname = 'uwnd';
time = 'all'; % Equivalently, one may use tim = [1:576], unless
% more months are added.
latlonlim = [0 360 -90 90];
level = 1;
[uwnd, lat, lon, time] = ...
getnc2(filename, varname, latlonlim, level, 'all');
% Make 4x6 degree averages
[ntim, nlat, nlon] = size(uwnd);
uwnd2 = repmat(NaN, [ntim nlat/2 nlon/3]);
for i = 1:nlat/2;
latind = 2*(i-1)+[1:2];
for j = 1:nlon/3;
lonind = 3*(j-1)+[1:3];
uwnd2(:,i,j) = ...
squeeze(mean2(mean2(shiftdim(uwnd(:,latind,lonind), 1))));
end
end
lat2 = repmat(NaN, [nlat/3 1]);
for i = 1:nlat/2;
latind = 2*(i-1)+[1:2];
lat2(i) = mean(lat(latind));
end
lon2 = repmat(NaN, [nlon/3 1]);
for j = 1:nlon/3;
lonind = 3*(j-1)+[1:3];
lon2(j) = mean(lon(lonind));
end
% Now, remove climatology, and rename uwnd2 to uwnd
[uwnd, uwndclim] = annave(uwnd2);
% Write uwnd and uwndclim to a netcdf file
% Define (open) NetCDF files
cd /home/disk/tao/data/coads/sstanom4by6
nc_in = netcdf('slpcoadsanom4by6.18541995.nc', 'nowrite');
nc_out = netcdf('uwndcoadsanom4by6.19501997.nc', 'write');
% Define other dimensions (lat, lon)
nc_out('lat') = length(lat2);
nc_out('lon') = length(lon2);
% Define record dimension (time)
nc_out('time') = 0; d = nc_out('time');
d(:) = length(time);
% Copy variable information (but not data) from nc_in to nc_out
nc_out < nc_in{'lat'};
nc_out{'lat'} < att(nc_in{'lat'});
nc_out < nc_in{'lon'};
nc_out{'lon'} < att(nc_in{'lon'});
nc_out < nc_in{'time'};
nc_out{'time'} < att(nc_in{'time'});
nc_out < nc_in{'data'};
nc_out{'data'} < att(nc_in{'data'});
% Store dimension variable data into nc_out
nc_out{'lat'}(:) = lat2;
nc_out{'lon'}(:) = lon2;
nc_out{'time'}(1:length(time)) = time;
nc_out{'time'}.units(:) = 'days since 1-1-1 00:00:0.0';
% Store uwnd data into nc_out
mv = nc_in{'data'}.missing_value(:);
ao = 0.0;
sf = 0.01;
uwnd2 = (uwnd-ao)/sf;
uwnd2(isnan(uwnd2)) = mv;
nc_out{'data'}(:,:,:) = uwnd2(:,:,:);
% Add some descriptive attributes
nc_out{'data'}.missing_value = ncshort(mv);
nc_out{'data'}.add_offset = ncfloat(ao);
nc_out{'data'}.scale_factor = ncfloat(sf);
nc_out{'data'}.units(:) = 'm/s';
nc_out{'data'}.long_name(:) = ...
'u-wind Monthly Mean at Surface';
nc_out{'data'}.dataset = 'COADS 2-degree Enhanced';
nc_out{'data'}.averaging = ...
'2 grid points in latitude, 3 grid points in longitude';
close(nc_out);
close(nc_in);
%%%%%%%%%%%%%%%%%%%%%%%%%%%% Next, UWND 1854--1995 %%%%%%%%%%%%%%
% Load SST data
cd /home/disk/tao/data/coads/coads1
filename = 'uwnd.mean.nc';
varname = 'uwnd';
tim1 = [1:12*(1949-1854+1)];
latlonlim = [0 360 -90 90];
level = 1;
[uwnd1, lat, lon, time1] = ...
getnc2(filename, varname, latlonlim, level, tim1);
cd /home/disk/tao/data/coads/coads1a
filename = 'uwnd.mean.5097.nc';
tim2 = [1:12*(1995-1950+1)];
[uwnd2, lat, lon, time2] = ...
getnc2(filename, varname, latlonlim, level, tim2);
uwnd = [uwnd1; uwnd2];
clear uwnd1 uwnd2;
% Make 4x6 degree averages
[ntim, nlat, nlon] = size(uwnd);
uwnd2 = repmat(NaN, [ntim nlat/2 nlon/3]);
for i = 1:nlat/2;
latind = 2*(i-1)+[1:2];
for j = 1:nlon/3;
lonind = 3*(j-1)+[1:3];
uwnd2(:,i,j) = ...
squeeze(mean2(mean2(shiftdim(uwnd(:,latind,lonind), 1))));
end
end
lat2 = repmat(NaN, [nlat/3 1]);
for i = 1:nlat/2;
latind = 2*(i-1)+[1:2];
lat2(i) = mean(lat(latind));
end
lon2 = repmat(NaN, [nlon/3 1]);
for j = 1:nlon/3;
lonind = 3*(j-1)+[1:3];
lon2(j) = mean(lon(lonind));
end
% Now, remove climatology, and rename uwnd2 to uwnd
[uwnd, uwndclim] = annave(uwnd2);
% Write uwnd and uwndclim to a netcdf file
% Define (open) NetCDF files
cd /home/disk/tao/data/coads/sstanom4by6
nc_in = netcdf('slpcoadsanom4by6.18541995.nc', 'nowrite');
nc_out = netcdf('uwndcoadsanom4by6.18541995.nc', 'clobber');
% Get time index
time = nc_in{'time', 1}(:);
% Define other dimensions (lat, lon)
nc_out('lat') = length(lat2);
nc_out('lon') = length(lon2);
% Define record dimension (time)
nc_out('time') = 0; d = nc_out('time');
d(:) = length(time);
% Copy variable information (but not data) from nc_in to nc_out
nc_out < nc_in{'lat'};
nc_out{'lat'} < att(nc_in{'lat'});
nc_out < nc_in{'lon'};
nc_out{'lon'} < att(nc_in{'lon'});
nc_out < nc_in{'time'};
nc_out{'time'} < att(nc_in{'time'});
nc_out < nc_in{'data'};
nc_out{'data'} < att(nc_in{'data'});
% Store dimension variable data into nc_out
nc_out{'lat'}(:) = lat2;
nc_out{'lon'}(:) = lon2;
nc_out{'time'}(1:length(time)) = time;
% Store uwnd data into nc_out
mv = nc_in{'data'}.missing_value(:);
ao = 0.0;
sf = 0.01;
uwnd2 = (uwnd-ao)/sf;
uwnd2(isnan(uwnd2)) = mv;
nc_out{'data'}(:,:,:) = uwnd2(:,:,:);
% Add some descriptive attributes
nc_out{'data'}.missing_value = ncshort(mv);
nc_out{'data'}.add_offset = ncfloat(ao);
nc_out{'data'}.scale_factor = ncfloat(sf);
nc_out{'data'}.units(:) = 'm/s';
nc_out{'data'}.long_name(:) = ...
'u-wind Monthly Mean at Surface';
close(nc_out);
close(nc_in);