sots_ncp_3_extractor.m

% Extracts the data from the relevant netcdf files

% Long term, could all the sots data be compiled into one netcdf file, and
% then selected by a deployment flag? In the meantime, use a cruder
% solution

function [mooring_data] = sots_ncp_extractor(deployment)

    if deployment == 'pulse7'

        % Opens the relevant netcdf file

        ncid = netcdf.open('IMOS_DWM-SOTS_REFOBKGTPCS_20100817_Pulse_FV02_Pulse-7-2010-Gridded-Data_END-20110708_C-20190522.nc');

        % In order to access a summary of the information contained in the netcdf
        % file, you can use the following command
        % ncdisp('IMOS_DWM-SOTS_REFOBKGTPCS_20100817_Pulse_FV02_Pulse-7-2010-Gridded-Data_END-20110708_C-20190522.nc')

        % We create a cell object containing the names of each of the variables
        % stored in the netcdf. Note that variables in a NetCDF file are indexed
        % from 0.

        varnames = cell(size(netcdf.inqVarIDs(ncid)));

        for i=0:(length(netcdf.inqVarIDs(ncid))-1)

            varnames{i+1} = netcdf.inqVar(ncid,i);

        end

        % We create and populate a Matalb structure named "pulse7", containing     
        % fields named for each of the variables of the NetCDF file, and subfields
        % containing the instrument name and deployment depth, where relevant.
        %
        % For example,          pulse7.DOX2.Aanderaa_Optode3975C_38_5m

        for i=0:length(varnames)-1

            try
                depths_dummy = netcdf.getAtt(ncid,i,'sensor_depth');

                depths_dummy = split(depths_dummy,';');

                depths_dummy = strtrim(depths_dummy);

                depths_dummy = strcat(depths_dummy,'m');

                sensor_dummy = netcdf.getAtt(ncid,i,'sensor_name');

                sensor_dummy = split(sensor_dummy,';');

                sensor_dummy = strtrim(sensor_dummy);

                serial_dummy = netcdf.getAtt(ncid,i,'sensor_serial_number');

                serial_dummy = split(serial_dummy,';');

                serial_dummy = strtrim(serial_dummy);

                names_dummy = matlab.lang.makeValidName(strcat(sensor_dummy,'_',serial_dummy,'_',depths_dummy));

                dummy_variable = netcdf.getVar(ncid,i);

                for j=1:length(names_dummy)

                    pulse7.(varnames{i+1}).(names_dummy{j}) = dummy_variable(:,j);

                end


            catch

                pulse7.(varnames{i+1}) = netcdf.getVar(ncid,i);

            end

        end

        % We transpose the time vector, as it has a different format.

        pulse7.TIME = pulse7.TIME;

        % Here we set a time offset, as MATLAB's serial date number format is based
        % on time since year 0, while the netcdf data's serial date number is from
        % 01/01/1950 00:00:00 UTC

        time_offset = datenum(1950,1,1,0,0,0);

        % We apply this time offset to the TIME variable of the structure, so that
        % data is now correctly timestamped in Matlab.

        pulse7.TIME = pulse7.TIME + time_offset.*ones(size(pulse7.TIME));

        % We close the netcdf file

        netcdf.close(ncid);

        % We now assign the relevant data from the extraction to the outputs of
        % the function

        mooring_data.time = pulse7.TIME;

        mooring_data.temp_C = pulse7.TEMP.Sea_BirdElectronics_SBE16plusV2_6331_31_1m;

        mooring_data.temp_C_qc = pulse7.TEMP_quality_control(:,1);

        mooring_data.psal_PSU = pulse7.PSAL.Sea_BirdElectronics_SBE16plusV2_6331_31_1m;

        mooring_data.psal_PSU_qc = pulse7.PSAL_quality_control(:,1);

        mooring_data.density_kgm3 = pulse7.DENSITY.Sea_BirdElectronics_SBE16plusV2_6331_31_1m;

        mooring_data.density_kgm3_qc = pulse7.DENSITY_quality_control(:,1);

        % DOX2.Aanderaa_Optode3975_1161_31_1m or Sea_BirdElectronics_SBE43_431635_31_1m
        mooring_data.dox2_umolkg = pulse7.DOX2.Aanderaa_Optode3975_1161_31_1m; 

        mooring_data.dox2_umolkg_qc = pulse7.DOX2_quality_control(:,1);

        mooring_data.dox2_sol_umolkg = pulse7.OXSOL.Sea_BirdElectronics_SBE16plusV2_6331_31_1m;

        mooring_data.dox2_sol_umolkg_qc = pulse7.OXSOL_quality_control(:,1);

        mooring_data.gastension_Pa = pulse7.TOTAL_GAS_PRESSURE.ProOceanus_GTD_29_102_15_31_1m*100; % Record is in hPa

        mooring_data.gastension_Pa_qc = pulse7.TOTAL_GAS_PRESSURE_quality_control;

        mooring_data.mld_m = pulse7.MLD;

        mooring_data.mld_m_qc = pulse7.MLD_quality_control;

        % Use Eric's simulated wind speeds (Eric Schulz BOM 10m winds)
        load('Pulse_7_Eric_wind_sim.mat','wind_sim_total');

        mooring_data.windspeed_ms = wind_sim_total;

    

    elseif deployment == 'pulse9'

        % Here the NetCDF file containing the Pulse 9 data is imported into
        % Matlab. This file must be in Matlab's current folder.

        ncid = netcdf.open('IMOS_ABOS-SOTS_RWBKGOTPCS_20120619_Pulse_FV02_Pulse-9-2012-Gridded-Data_END-20130510_C-20210131.nc');


        % In order to access a summary of the information contained in the netcdf
        % file, you can use the following command
        % ncdisp('IMOS_ABOS-SOTS_RWBKGOTPCS_20120619_Pulse_FV02_Pulse-9-2012-Gridded-Data_END-20130510_C-20190910.nc')

        % We create a cell object containing the names of each of the variables
        % stored in the netcdf. Note that variables in a NetCDF file are indexed
        % from 0.

        varnames = cell(size(netcdf.inqVarIDs(ncid)));

        for i=0:(length(netcdf.inqVarIDs(ncid))-1)

            varnames{i+1} = netcdf.inqVar(ncid,i);

        end

        % We create and populate a Matalb structure named "pulse9", containing     
        % fields named for each of the variables of the NetCDF file, and subfields
        % containing the instrument name and deployment depth, where relevant.
        %
        % For example,          pulse9.DOX2.Aanderaa_Optode3975C_38_5m

        for i=0:length(varnames)-1

            try
                depths_dummy = netcdf.getAtt(ncid,i,'sensor_depth');

                depths_dummy = split(depths_dummy,';');

                depths_dummy = strtrim(depths_dummy);

                depths_dummy = strcat(depths_dummy,'m');

                sensor_dummy = netcdf.getAtt(ncid,i,'sensor_name');

                sensor_dummy = split(sensor_dummy,';');

                sensor_dummy = strtrim(sensor_dummy);

                names_dummy = matlab.lang.makeValidName(strcat(sensor_dummy,'_',depths_dummy));

                dummy_variable = netcdf.getVar(ncid,i);

                for j=1:length(names_dummy)

                    pulse9.(varnames{i+1}).(names_dummy{j}) = dummy_variable(:,j);

                end


            catch

                dummy_variable = netcdf.getVar(ncid,i);

                if length(dummy_variable) > 100

                    pulse9.(varnames{i+1}) = dummy_variable; 

                else

                    pulse9.(varnames{i+1}) = netcdf.getVar(ncid,i);

                end

            end

        end

        % We transpose the time vector, as it has a different format.

        pulse9.TIME = pulse9.TIME;

        % Here we set a time offset, as MATLAB's serial date number format is based
        % on time since year 0, while the netcdf data's serial date number is from
        % 01/01/1950 00:00:00 UTC

        time_offset = datenum(1950,1,1,0,0,0);

        % We apply this time offset to the TIME variable of the structure, so that
        % data is now correctly timestamped in Matlab.

        pulse9.TIME = pulse9.TIME + time_offset.*ones(size(pulse9.TIME));

        % We adjust the VEMCO temperature sensors from 45-85m in order for them to
        % read in agreement with the two Seabird sensors they sit between, by
        % applying linear offsets.

        vemco_offsets = [0.025 0.025 0.015 0.045 0.015 -0.02 0.03 0];

        fields_dummy=fieldnames(pulse9.TEMP);

        for i=1:length(vemco_offsets)

            pulse9.TEMP.(fields_dummy{i+1}) = pulse9.TEMP.(fields_dummy{i+1}) + vemco_offsets(i);

        end

        % We close the netcdf file

        netcdf.close(ncid);

        % We import the matching sofs3 surface data into Matlab, to access
        % windspeed data.

        ncid2 = netcdf.open('IMOS_DWM-ASFS_CFMST_20120714T080000Z_SOFS_FV02_SOFS-3-2012_END-20130102T000000Z_C-20180716T064518Z.nc');

        % We perform the same routine as with the Pulse 9 data, to extract the
        % data and place it in a structure named "sofs3".

        varnames2 = cell(size(netcdf.inqVarIDs(ncid2)));

        for i=0:(length(netcdf.inqVarIDs(ncid2))-1)

            varnames2{i+1} = netcdf.inqVar(ncid2,i);

        end

        for i=0:length(varnames2)-1

            try
                depths_dummy = netcdf.getAtt(ncid2,i,'sensor_depth');

                depths_dummy = split(depths_dummy,';');

                depths_dummy = strtrim(depths_dummy);

                depths_dummy = strcat(depths_dummy,'m');

                sensor_dummy = netcdf.getAtt(ncid2,i,'sensor_name');

                sensor_dummy = split(sensor_dummy,';');

                sensor_dummy = strtrim(sensor_dummy);

                names_dummy = matlab.lang.makeValidName(strcat(sensor_dummy,'_',depths_dummy));

                dummy_variable = netcdf.getVar(ncid2,i);

                for j=1:length(names_dummy)

                    sofs3.(varnames2{i+1}).(names_dummy{j}) = dummy_variable(:,j);

                end


            catch

                sofs3.(varnames2{i+1}) = netcdf.getVar(ncid2,i);

            end

        end

        sofs3.TIME = sofs3.TIME + time_offset.*ones(size(sofs3.TIME));

        % As sofs3 surface data has been recorded in minute intervals, from a
        % different starting date to Pulse 10, here we create a time index to
        % subsample the sofs3 data to align it with Pulse 9 sampling.

        % This results in the following being true:
        % sofs3.TIME(sofs3timeindex) = pulse9.TIME

        sofs3timeindex = zeros(size(pulse9.TIME));

        for i = 1:length(pulse9.TIME)

            try

                sofs3timeindex(i) = find(sofs3.TIME == pulse9.TIME(i),1);

            catch

                sofs3timeindex(i) = 1;

            end

        end

        % We close the netcdf
        netcdf.close(ncid2);

        % We now assign the relevant data from the extraction to the outputs of
        % the function

        mooring_data.time = pulse9.TIME;

        mooring_data.temp_C = pulse9.TEMP.Sea_BirdElectronics_SBE16plusV2_28_5m;

        mooring_data.temp_C_qc = pulse9.TEMP_quality_control(:,1);

        mooring_data.psal_PSU = pulse9.PSAL.Sea_BirdElectronics_SBE16plusV2_28_5m;

        mooring_data.psal_PSU_qc = pulse9.PSAL_quality_control(:,1);

        mooring_data.density_kgm3 = pulse9.DENSITY.Sea_BirdElectronics_SBE16plusV2_28_5m;

        mooring_data.density_kgm3_qc = pulse9.DENSITY_quality_control(:,1);

        % Aanderaa_Optode3975C_38_5m or Sea_BirdElectronics_SBE43_38_5m
        mooring_data.dox2_umolkg = pulse9.DOX2.Aanderaa_Optode3975C_28_5m; 

        mooring_data.dox2_umolkg_qc = pulse9.DOX2_quality_control(:,1);

        mooring_data.dox2_sol_umolkg = pulse9.OXSOL.Sea_BirdElectronics_SBE16plusV2_28_5m;

        mooring_data.dox2_sol_umolkg_qc = pulse9.OXSOL_quality_control(:,1);

        mooring_data.gastension_Pa = pulse9.TOTAL_GAS_PRESSURE.ProOceanus_GTD_28_5m/(1E3); % Record is in millibars in netcdf

        mooring_data.gastension_Pa_qc = pulse9.TOTAL_GAS_PRESSURE_quality_control;

        mooring_data.mld_m = pulse9.MLD;

        mooring_data.mld_m_qc = pulse9.MLD_quality_control;

        mooring_data.windspeed_ms = sofs3.WSPD10M(sofs3timeindex);

        % No atmospheric pressure available

    

    elseif deployment == 'sofs75'


        %     addpath('C:\xx\xx\data_folder\composite1.dat')
        %     
        %     addpath('C:\xx\xx\data_folder\composite1.dat')

        % Here the NetCDF file containing the SOFS-7.5 data is imported into
        % Matlab. This file must be in Matlab's current folder.

        ncid = netcdf.open('IMOS_ABOS-SOTS_FRVMWEBUOSTCGP_20180628_SOFS_FV02_SOFS-7.5-2018-Gridded-Data_END-20190423_C-20190916.nc');

        % In order to access a summary of the information contained in the netcdf
        % file, you can use the following command
        % ncdisp('IMOS_ABOS-SOTS_FRVMWEBUOSTCGP_20180628_SOFS_FV02_SOFS-7.5-2018-Gridded-Data_END-20190423_C-20190916.nc')

        % We create a cell object containing the names of each of the variables
        % stored in the netcdf. Note that variables in a NetCDF file are indexed
        % from 0.

        varnames = cell(size(netcdf.inqVarIDs(ncid)));

        for i=0:(length(netcdf.inqVarIDs(ncid))-1)

            varnames{i+1} = netcdf.inqVar(ncid,i);

        end

        % We create and populate a Matalb structure named "sofs75", containing     
        % fields named for each of the variables of the NetCDF file, and subfields
        % containing the instrument name and deployment depth, where relevant.
        %
        % For example,          sofs75.DOX2.Aanderaa_Optode3975C_38_5m

        for i=0:length(varnames)-1

            try
                depths_dummy = netcdf.getAtt(ncid,i,'sensor_depth');

                depths_dummy = split(depths_dummy,';');

                depths_dummy = strtrim(depths_dummy);

                depths_dummy = strcat(depths_dummy,'m');

                sensor_dummy = netcdf.getAtt(ncid,i,'sensor_name');

                sensor_dummy = split(sensor_dummy,';');

                sensor_dummy = strtrim(sensor_dummy);

                serial_dummy = netcdf.getAtt(ncid,i,'sensor_serial_number');

                serial_dummy = split(serial_dummy,';');

                serial_dummy = strtrim(serial_dummy);

                names_dummy = matlab.lang.makeValidName(strcat(sensor_dummy,'_',serial_dummy,'_',depths_dummy));

                dummy_variable = netcdf.getVar(ncid,i);

                for j=1:length(names_dummy)

                    sofs75.(varnames{i+1}).(names_dummy{j}) = dummy_variable(:,j);

                end


            catch

                sofs75.(varnames{i+1}) = netcdf.getVar(ncid,i);

            end

        end

        % We extract the global attributes from the netcdf file and insert them
        % into the structure



        % We transpose the time vector, as it has a different format.

        sofs75.TIME = sofs75.TIME;

        % Here we set a time offset, as MATLAB's serial date number format is based
        % on time since year 0, while the netcdf data's serial date number is from
        % 01/01/1950 00:00:00 UTC

        time_offset = datenum(1950,1,1,0,0,0);

        % We apply this time offset to the TIME variable of the structure, so that
        % data is now correctly timestamped in Matlab.

        sofs75.TIME = sofs75.TIME + time_offset.*ones(size(sofs75.TIME));

        % We close the netcdf file

        netcdf.close(ncid);


        % We now assign the relevant data from the extraction to the outputs of
        % the function

        mooring_data.time = sofs75.TIME;

        mooring_data.temp_C = sofs75.TEMP.Sea_BirdElectronics_SBE37SMP_ODO_15696_30_0m;

        mooring_data.temp_C_qc = sofs75.TEMP_quality_control(:,6);

        mooring_data.psal_PSU = sofs75.PSAL.Sea_BirdElectronics_SBE37SMP_ODO_15696_30_0m;

        mooring_data.psal_PSU_qc = sofs75.PSAL_quality_control(:,4);

        % Need to manually calculate density!

        SP = mooring_data.psal_PSU;

        t = mooring_data.temp_C;

        p = sofs75.PRES.Sea_BirdElectronics_SBE37SMP_ODO_15696_30_0m;

        long = sofs75.LONGITUDE;

        lat = sofs75.LATITUDE;

        SA = gsw_SA_from_SP(SP,p,long,lat);

        CT = gsw_CT_from_t(SA,t,p);

        mooring_data.density_kgm3 = gsw_rho(SA,CT,p);

        mooring_data.density_kgm3_qc = max([mooring_data.psal_PSU_qc,mooring_data.temp_C_qc,sofs75.PRES_quality_control(:,1)],[],2);

        mooring_data.dox2_umolkg = sofs75.DOX2.Sea_BirdElectronics_SBE37SMP_ODO_15696_30_0m; 

        mooring_data.dox2_umolkg_qc = sofs75.DOX2_quality_control(:,2);

        mooring_data.dox2_sol_umolkg = sofs75.OXSOL.Sea_BirdElectronics_SBE37SMP_ODO_15696_30_0m;

        mooring_data.dox2_sol_umolkg_qc = sofs75.OXSOL_quality_control(:,2);
        
        mooring_data.sub_mld_dox2_umolkg = sofs75.DOX2.Sea_BirdElectronics_SBE37SMP_ODO_15972_480_0m;
        
        mooring_data.sub_mld_dox2_umolkg_qc = sofs75.DOX2_quality_control(:,end);

        mooring_data.gastension_Pa = sofs75.TOTAL_GAS_PRESSURE.ProOceanus_TGTD_37_468_33_30_0m*100; % Record is in hPa in netcdf

        mooring_data.gastension_Pa_qc = sofs75.TOTAL_GAS_PRESSURE_quality_control(:,2);

        % Build mld from FV02 file
        build_mld = true;

        % Calculate sensor offsets from FV02 file
        manual_offset = true;

        if build_mld

            if manual_offset

                % If temperature offsets haven't been applied in the FV02 file
                % already
                sofs75.TEMParray = struct2array(sofs75.TEMP);

                % We select the profiles where all instruments record QC values of 1

                sofs75.valid_record_start = find(all(sofs75.TEMP_quality_control==1,2),1);

                sofs75.valid_record_end = find(all(sofs75.TEMP_quality_control==1,2),1,'last');

                sofs75.TEMParray_qc1_profiles = sofs75.TEMParray(sofs75.valid_record_start:sofs75.valid_record_end,:);

                sofs75.TIME_qc1_profiles = sofs75.TIME(sofs75.valid_record_start:sofs75.valid_record_end);

                % We create an array of Nan the same size as 'sofs75.TEMParray_qc1_profiles', to
                % store the excursions of the Starmon sensor values from the Seabird sensor
                % values.

                sofs75.TEMPexcursions = NaN(size(sofs75.TEMParray_qc1_profiles));

                % We set a temperature threshold of 0.004°C, based on the accuracy of
                % +-0.002°C of the SBE 37-SMP-ODO, as in a body of water of one temperature,
                % two SBE 37s could read 0.004°C out from each other.

                temp_thresh = 0.004;

                % We give the indices of the trusted temperature sources (generally SBEs)

                sofs75.trusted_temps = [1 6 14 18 25];

                % We loop through all the selected temperature profiles

                for i = 1:length(sofs75.TEMPexcursions)

                        % We create multiple if conditions. Each condition checks for the current 
                        % time stamp if the pairs of Seabird sensors closest to each other 
                        % agree to within the temp_thresh. If they do, the excursions of
                        % the Starmon sensors in between the two Seabirds are recorded in
                        % the array 'sofs75.TEMPexcursions' for the relevant time stamp.

                    for j=1:(length(sofs75.trusted_temps)-1)

                        if abs(sofs75.TEMParray_qc1_profiles(i,sofs75.trusted_temps(j))-sofs75.TEMParray_qc1_profiles(i,sofs75.trusted_temps(j+1)))<=temp_thresh

                            sofs75.TEMPexcursions(i,(sofs75.trusted_temps(j)+1):(sofs75.trusted_temps(j+1)-1)) = mean([sofs75.TEMParray_qc1_profiles(i,sofs75.trusted_temps(j)) sofs75.TEMParray_qc1_profiles(i,sofs75.trusted_temps(j+1))])  - sofs75.TEMParray_qc1_profiles(i,(sofs75.trusted_temps(j)+1):(sofs75.trusted_temps(j+1)-1));

                        end

                    end

                end
                % We calculate the median value of the offsets for each temperature sensor,
                % ignoring any NaN values generated by the Seabird sensors.

                sofs75.TEMPoffsets = median(sofs75.TEMPexcursions,'omitnan');

                % We set the NaN results from the Seabirds to 0.

                sofs75.TEMPoffsets(isnan(sofs75.TEMPoffsets))=0;

                % We adjust the selected temperature readings in sofs75.TEMParray_qc1_profiles 
                % and sofs75.TEMParray by the calculated offsets.

                sofs75.TEMParray_qc1_profiles = sofs75.TEMParray_qc1_profiles + sofs75.TEMPoffsets;

                sofs75.TEMParray = sofs75.TEMParray + sofs75.TEMPoffsets;

            else

                sofs75.TEMParray = struct2array(sofs75.TEMP);

            end

            % *********************************************************************** %
            % ----------------------------------------------------------------------- %
            %                Calculation of Mixed layer depth                         %
            % ----------------------------------------------------------------------- %
            % *********************************************************************** %

            % We calculate the mixed layer depth using a threshold method. 

            % First a temperature threshold is set.

            MLDP_temp_thresh = 0.2;

            % We create a new field in the sofs75 structure to store the MLDP data.

            sofs75.MLDP.thresh02 = max(sofs75.DEPTH_TEMP).*ones(length(sofs75.TEMParray),1);

            % For each timestamp, the MLDP is set at the depth of the first sensor whose
            % absolute difference from the shallowest sensor exceeds the threshold. If
            % no sensor exceeds this threshold, then the MLDP is set as the deepest
            % sensor depth.

            for i = 1:length(sofs75.MLDP.thresh02)

                if any(abs(sofs75.TEMParray(i,1)-sofs75.TEMParray(i,:))>=MLDP_temp_thresh)

                    sofs75.MLDP.thresh02(i) = sofs75.DEPTH_TEMP(find(abs(sofs75.TEMParray(i,1)-sofs75.TEMParray(i,:))>=MLDP_temp_thresh,1));

                end

            end

            % Here we use a linear interpolation method, as suggested in Huang et al (2018)
            % as the most accurate method for temperature data, spare in depth resolution.

            sofs75.MLDP.interp02 = nan.*ones(length(sofs75.TEMParray),1);

            MLDP_temp_interp02 = 0.2;

            for i = 1:length(sofs75.MLDP.interp02)

                if sum(isnan(sofs75.TEMParray(i,:)))==0 && ~any(abs(sofs75.TEMParray(i,1)-sofs75.TEMParray(i,:))>=MLDP_temp_interp02)

                    sofs75.MLDP.interp02(i) = max(sofs75.DEPTH_TEMP);   

                elseif sum(isnan(sofs75.TEMParray(i,:)))==0

                    idx_below = find(abs(sofs75.TEMParray(i,1)-sofs75.TEMParray(i,:))>=MLDP_temp_interp02,1);

                    idx_above = idx_below-1;

                    sofs75.MLDP.interp02(i) = interp1(sofs75.TEMParray(i,idx_above:idx_below),sofs75.DEPTH_TEMP(idx_above:idx_below),sofs75.TEMParray(i,1)-MLDP_temp_interp02);        

                end

            end


        end

        mooring_data.mld_m = sofs75.MLDP.interp02;

        mooring_data.mld_m_qc = max(sofs75.TEMP_quality_control,[],2);

        % Windspeed measured at 2.61m, scale using the power law from
        % Justus and Mikhail (1976)
        
        for w_idx=1:length(sofs75.WSPD(:,1))
            
            wind_alpha(w_idx) = (0.37-0.088*log(sofs75.WSPD(w_idx,1)))/(1-0.088*log(2.61/10));
            
            windspeed_10m(w_idx) = sofs75.WSPD(w_idx,1) * (10/2.61)^wind_alpha(w_idx);
            
        end
        
%         % Currently using linearly interpolated 10m windspeeds from the 2
%         % hour realtime SOTS files, downloaded from the AODN as of
%         % 14/05/20, will progress to using the 10m windspeeds from FV02
%         % file once available
%         load('sofs75_wspd10_interp.mat','WSPD10M_interp')
%         
%         mooring_data.windspeed_ms = WSPD10M_interp;

        mooring_data.windspeed_ms = sofs75.WSPD(:,1);

        mooring_data.windspeed_ms_qc = sofs75.WSPD_quality_control(:,1);

        mooring_data.atmosphericpress_Pa = sofs75.CAPH(:,1)*100; % Record is in hPa

    end
    
end