source: TOOLS/CMIP6_FORCING/SOLAR/CMIP6_sensitivity_2bands.m @ 3383

% MATLAB run script to integrate spectral solar irradiance (SSI) data
% recommended and provided by SPARC/SOLARIS-HEPPA for use in CMIP6 to
% spectral bands of any desired resolution (according to
% radiation/photolysis scheme of employed (chemistry) climate model)
%
% Author: Tim Kruschke (tkruschke[at]geomar.de)
% Last modified: 02 May 2016 (compliance with solarforcing_*_*_3.1.nc)
%
% Description:
%    1. reading data from netcdf-file as provided by SPARC/SOLARIS-HEPPA
%    2. reading definitions of spectral bands (user-specified by creating an
%       ascii-textfile according to specific formatting, as described below)
%    3. calling MATLAB-function "regrid_ssi" (provided with this run script as
%       part of a joint zip-archive) to integrate SSI over specified bands
%    4. writing new netcdf-file with same structure and attributes as original
%       (input) file, but only with (unchanged) total solar irradiance
%       (TSI), SSI integrated over spectral bands specified by user, and
%       irradiance fraction (of TSI) in respective band (plus
%       necessary/related dimensions/variables time, wlen, wlenbinsize)
%
% Formatting of ascii-textfile, containing definitions of spectral bands:
%    NO HEADER LINES, NO UNITS, NO CHARACTER DELIMITERS
%    It's as simple as this:
%    - one line per spectral band
%    - each line contains two floating point numbers: lower and upper edge
%      of respective spectral band (in nanometers!!!), separated by
%      blank space(s) or tabulator(s)
%    See example_spectral_bands.dat (provided with this run script as
%    part of a joint zip-archive) as an example.
%--------------------------------------------------------------------------

%  execute with run CMIP6.m

%% User-defined paths and filenames

% path to directory of input data; adapt to your file structure
dir_in=['.'];

% path to directory of output data; adapt to your file structure if desired
% to differ from input directory
dir_out=dir_in;

% filename of ascii-textfile (to be placed in dir_in) containing
% definitions of spectral bands (see explanations in script header);
% adapt to name of your band definition file
%filename_spectral_bands='example_spectral_bands.dat';
%filename_spectral_bands='example_spectral_bands_6_SSI_FRAC=1.dat';
filename_spectral_bands='example_spectral_bands_2_SSI_FRAC=1.dat';

% filename of original data file, provided by SPARC/SOLARIS-HEPPA (to be
% placed in dir_in); adapt to respective file that is to be considered
%filename_in='solarforcing_ext_mon_3.1.nc';
filename_in='/prodigfs/project/input4MIPs/SOLAR/sensitivity/solarforcing_ext_day_3.1.nc';

% lines 53-54 only used to match input and output filenames
% ATTENTION: works only on unix platforms; if run on other platforms,
% delete/comment following two lines and enter complete filename_out in
% l. 59
[~,filename_in_base]=unix(['basename ' dir_in filesep filename_in ' .nc']);
filename_in_base=strcat(filename_in_base);

% filename of new file to be created, containing SSI data in spectral
% resolutiuon as desired by the user; adapt to whatever name you like if
% not desired to match input filename
%filename_out=[filename_in_base '_SSI_regridded.nc'];
filename_out=[filename_in_base '_SSI_regridded_2_TSI_anomaly.nc'];

%% Usually, nothing should be edited below this line!!!!!!!!!!!!!!!!!!!!!!!

ncid_in=netcdf.open([dir_in filesep filename_in],'NOWRITE');
ncid_out=netcdf.create(strcat([dir_out filesep filename_out]),'CLOBBER'); % existing files are overwritten

% adapting title of original file
nc_title=netcdf.getAtt(ncid_in,-1,'title');
nc_title=['Spectral solar irradiance (SSI) of ' nc_title ', converted to custom spectral resolution'];
netcdf.putAtt(ncid_out,-1,'title',nc_title);

%% time stamp
netcdf.putAtt(ncid_out,-1,'creation_date',datestr(now)); % creates a time stamp, though not compliant with ISO 8601
%c% You can use the following lines (commented out) instead to generate a
%c% time stamp compliant to ISO 8601; change last digits according to
%c% your time zone in this case (currently +02:00 for CEST)
%     create_date_vec=datevec(now);
%     create_date_str=cell(1,6);
%     for id=1:6;
%         if create_date_vec(id)<10;
%             create_date_str{1,id}=['0' num2str(round(create_date_vec(id)))];
%         else
%             create_date_str{1,id}=num2str(round(create_date_vec(id)));
%         end
%     end
%     netcdf.putAtt(ncid_out,-1,'creation_date',[create_date_str{1,1} '-' create_date_str{1,2} '-' ...
%         create_date_str{1,3} 'T' create_date_str{1,4} ':' create_date_str{1,5} ':' create_date_str{1,6} '+02:00'])
%%

% copy global attributes to new file
netcdf.copyAtt(ncid_in,-1,'institution',ncid_out,-1)
netcdf.copyAtt(ncid_in,-1,'institution_id',ncid_out,-1)
netcdf.copyAtt(ncid_in,-1,'activity_id',ncid_out,-1)
netcdf.copyAtt(ncid_in,-1,'comment',ncid_out,-1)
netcdf.copyAtt(ncid_in,-1,'time_coverage_start',ncid_out,-1)
netcdf.copyAtt(ncid_in,-1,'time_coverage_end',ncid_out,-1)
netcdf.copyAtt(ncid_in,-1,'frequency',ncid_out,-1)
netcdf.copyAtt(ncid_in,-1,'source',ncid_out,-1)
netcdf.copyAtt(ncid_in,-1,'source_id',ncid_out,-1)
netcdf.copyAtt(ncid_in,-1,'realm',ncid_out,-1)
netcdf.copyAtt(ncid_in,-1,'further_info_url',ncid_out,-1)
netcdf.copyAtt(ncid_in,-1,'metadata_url',ncid_out,-1)
netcdf.copyAtt(ncid_in,-1,'contributor_name',ncid_out,-1)
netcdf.copyAtt(ncid_in,-1,'references',ncid_out,-1)
netcdf.copyAtt(ncid_in,-1,'Conventions',ncid_out,-1)
netcdf.putAtt(ncid_out,-1,'history',['Original SSI data from file ' filename_in ' integrated ' ...
    'over spectral bands as defined in file ' filename_spectral_bands ' ' ...
    'by MATLAB run script CMIP6_SSI_convert_resolution_run.m, using function regrid_ssi.m']);

% reading SSI-related variables from CMIP6 input file
wvl_in_ID=netcdf.inqVarID(ncid_in,'wlen');
wvl_edges_in_ID=netcdf.inqVarID(ncid_in,'wlen_bnds');
bw_in_ID=netcdf.inqVarID(ncid_in,'wlenbinsize');
time_in_ID=netcdf.inqVarID(ncid_in,'time');
time_bnds_in_ID=netcdf.inqVarID(ncid_in,'time_bnds');
year_in_ID=netcdf.inqVarID(ncid_in,'calyear');
month_in_ID=netcdf.inqVarID(ncid_in,'calmonth');
day_in_ID=netcdf.inqVarID(ncid_in,'calday');
tsi_in_ID=netcdf.inqVarID(ncid_in,'tsi');
ssi_in_ID=netcdf.inqVarID(ncid_in,'ssi');

wvl_in=netcdf.getVar(ncid_in,wvl_in_ID);
wvl_edges_in=(netcdf.getVar(ncid_in,wvl_edges_in_ID))';
bw_in=netcdf.getVar(ncid_in,bw_in_ID);
ssi_in=netcdf.getVar(ncid_in,ssi_in_ID);
tsi_in=netcdf.getVar(ncid_in,tsi_in_ID);
time_in=netcdf.getVar(ncid_in,time_in_ID);
time_bnds_in=netcdf.getVar(ncid_in,time_bnds_in_ID);
year_in=netcdf.getVar(ncid_in,year_in_ID);
month_in=netcdf.getVar(ncid_in,month_in_ID);
day_in=netcdf.getVar(ncid_in,day_in_ID);

% read user specified definitions of spectral bands
fid=fopen([dir_in filesep filename_spectral_bands]);
tmp=textscan(fid,'%f %f');
fclose(fid);
wvl_edges_out=tmp{1,1};
wvl_edges_out(:,2)=tmp{1,2};
clear tmp

% calculation of output central wave lengths
wvl_out=sum(wvl_edges_out,2)/2;

% calculation of output bin width
bw_out=wvl_edges_out(:,2)-wvl_edges_out(:,1);

% check for consistency of output bin definitions
if min(bw_out)<0;
    error(['Error: For at least one spectral band, the specified upper edge \n' ...
        '(right column) in your input file ' filename_in '\n ' ...
        'is smaller than the respective lower edge (left column).'])
end

% check for spectral coverage of input data with respect to desired output
% bins
if min(wvl_edges_out(:,1))<min(wvl_edges_in(:,1)) || max(wvl_edges_out(:,2))>max(wvl_edges_in(:,2));
    warning(['You are extrapolating into spectral ranges not covered by the input spectrum! ' ...
        'This routine supports this by linearly extrapolating based on change between last two bins. ' ...
        'However, this might be physically inappropriate!']);
end

% Integration of SSI data for each specified spectral band
% (regrid_ssi.m calculates weighted means of SSI per nm, this is multiplied by binwidth)
ssi_out=bsxfun(@times,regrid_ssi(ssi_in,wvl_edges_in,wvl_edges_out),bw_out);

% Calculation of solar irradiance fraction (compared to total solar
% irradiance) in each band
ssi_frac_out=bsxfun(@rdivide,ssi_out,tsi_in');

% Calculation the average value of tsi
avtsi=mean(tsi_in)
disp (avtsi)

%Calculate the anomaly of tsi
antsi(:,1)=tsi_in(:,1)-avtsi

disp (antsi)

format longEng
disp (avtsi)

% writing data (including dimensions) to output file
  % dimension definitions
time_dimID=netcdf.defDim(ncid_out,'time',0);
wvl_dimID=netcdf.defDim(ncid_out,'wlen',length(wvl_out));
bounds_dimID=netcdf.defDim(ncid_out,'nbd',2);

  % variable definitions
time_out_ID=netcdf.defVar(ncid_out,'time','double',time_dimID);
time_bnds_out_ID=netcdf.defVar(ncid_out,'time_bnds','double',[bounds_dimID time_dimID]);
wvl_out_ID=netcdf.defVar(ncid_out,'wlen','double',wvl_dimID);
wvl_edges_out_ID=netcdf.defVar(ncid_out,'wlen_bnds','double',[wvl_dimID bounds_dimID]);
bw_out_ID=netcdf.defVar(ncid_out,'wlenbinsize','double',wvl_dimID);
year_out_ID=netcdf.defVar(ncid_out,'calyear','double',time_dimID);
month_out_ID=netcdf.defVar(ncid_out,'calmonth','double',time_dimID);
day_out_ID=netcdf.defVar(ncid_out,'calday','double',time_dimID);
tsi_out_ID=netcdf.defVar(ncid_out,'tsi','float',time_dimID);
antsi_out_ID=netcdf.defVar(ncid_out,'antsi','float',time_dimID);
ssi_out_ID=netcdf.defVar(ncid_out,'ssi','float',[wvl_dimID time_dimID]);
ssi_frac_out_ID=netcdf.defVar(ncid_out,'ssi_frac','float',[wvl_dimID time_dimID]);

  % copy variable attributes from input to output file
netcdf.copyAtt(ncid_in,time_in_ID,'axis',ncid_out,time_out_ID)
netcdf.copyAtt(ncid_in,time_in_ID,'long_name',ncid_out,time_out_ID)
netcdf.copyAtt(ncid_in,time_in_ID,'standard_name',ncid_out,time_out_ID)
netcdf.copyAtt(ncid_in,time_in_ID,'units',ncid_out,time_out_ID)
netcdf.copyAtt(ncid_in,time_in_ID,'calendar',ncid_out,time_out_ID)
netcdf.copyAtt(ncid_in,time_in_ID,'_CoordinateAxisType',ncid_out,time_out_ID)
netcdf.copyAtt(ncid_in,time_in_ID,'bounds',ncid_out,time_out_ID)

netcdf.copyAtt(ncid_in,time_bnds_in_ID,'long_name',ncid_out,time_bnds_out_ID)
netcdf.copyAtt(ncid_in,time_bnds_in_ID,'units',ncid_out,time_bnds_out_ID)

netcdf.copyAtt(ncid_in,wvl_in_ID,'standard_name',ncid_out,wvl_out_ID)
netcdf.copyAtt(ncid_in,wvl_in_ID,'long_name',ncid_out,wvl_out_ID)
netcdf.copyAtt(ncid_in,wvl_in_ID,'units',ncid_out,wvl_out_ID)
netcdf.copyAtt(ncid_in,wvl_in_ID,'bounds',ncid_out,wvl_out_ID)

netcdf.copyAtt(ncid_out,wvl_edges_in_ID,'long_name',ncid_out,wvl_edges_out_ID)
netcdf.copyAtt(ncid_out,wvl_edges_in_ID,'units',ncid_out,wvl_edges_out_ID)

netcdf.copyAtt(ncid_in,bw_in_ID,'long_name',ncid_out,bw_out_ID)
netcdf.copyAtt(ncid_in,bw_in_ID,'units',ncid_out,bw_out_ID)

netcdf.copyAtt(ncid_in,year_in_ID,'long_name',ncid_out,year_out_ID)

netcdf.copyAtt(ncid_in,month_in_ID,'long_name',ncid_out,month_out_ID)

netcdf.copyAtt(ncid_in,day_in_ID,'long_name',ncid_out,day_out_ID)

netcdf.copyAtt(ncid_in,tsi_in_ID,'standard_name',ncid_out,tsi_out_ID)
netcdf.copyAtt(ncid_in,tsi_in_ID,'long_name',ncid_out,tsi_out_ID)
netcdf.copyAtt(ncid_in,tsi_in_ID,'units',ncid_out,tsi_out_ID)
netcdf.copyAtt(ncid_in,tsi_in_ID,'cell_methods',ncid_out,tsi_out_ID)

netcdf.putAtt(ncid_out,antsi_out_ID,'standard_name','total solar_irradiance anomaly')
netcdf.putAtt(ncid_out,antsi_out_ID,'long_name','reconstructed total solar irradiance anomaly at 1 AU')
netcdf.putAtt(ncid_out,antsi_out_ID,'units','W m^-2')

netcdf.putAtt(ncid_out,ssi_out_ID,'long_name','reconstructed solar irradiance at 1 AU within respective bands')
netcdf.putAtt(ncid_out,ssi_out_ID,'units','W m^-2')

netcdf.putAtt(ncid_out,ssi_frac_out_ID,'long_name','reconstructed fraction of total solar irradiance at 1 AU within respective bands')
netcdf.putAtt(ncid_out,ssi_frac_out_ID,'units','1')

netcdf.close(ncid_in);
netcdf.endDef(ncid_out);

netcdf.putVar(ncid_out,time_out_ID,0,length(time_in),time_in)
netcdf.putVar(ncid_out,time_bnds_out_ID,time_bnds_in)
netcdf.putVar(ncid_out,wvl_out_ID,wvl_out)
netcdf.putVar(ncid_out,wvl_edges_out_ID,wvl_edges_out)
netcdf.putVar(ncid_out,bw_out_ID,bw_out)
netcdf.putVar(ncid_out,year_out_ID,year_in)
netcdf.putVar(ncid_out,month_out_ID,month_in)
netcdf.putVar(ncid_out,day_out_ID,day_in)
netcdf.putVar(ncid_out,tsi_out_ID,tsi_in)
netcdf.putVar(ncid_out,antsi_out_ID,antsi)
netcdf.putVar(ncid_out,ssi_out_ID,ssi_out)
netcdf.putVar(ncid_out,ssi_frac_out_ID,ssi_frac_out)

netcdf.sync(ncid_out);
netcdf.close(ncid_out);