Get the weekly average of daily .netcdf data

I am using OSCAR 2.0 ocean current velocity data: https://podaac.jpl.nasa.gov/dataset/OSCAR_L4_OC_FINAL_V2.0#

Each daily file contains the following variables:

Lat (719 x 1 double), Long (1440 x 1 double), Time, Zonal Velocity "U" (1440 x 719 double), & Meridional Velocity "V" (1440 x 719 double).

I need to find the weekly average of the zonal and meridional velocity over 3 years: 2008-05-01 to 2001-04-31. Rather than having a bunch of daily files, I need weekly files. The lat, long variables will keep the same values. For the first 7 daily files, time = 1. The next 7 files, time = 2, etc. "Avg U" should be the average of the first seven "u".

Ex) File #1

Lat (719x1 double), Long (1440 x 1 double), Time = 1, AvgU (1440 x 719 double), Avg V (1440 x 719 double)

Here is what I've been working with, but clearly it is unfinished / wrong.

% have a bunch of daily files
% extract 7 files at a time 
% take the average of those 7 files 

days = 20080501 - 20110431
n = length(days) 
for i = 1:n
    name = int2str(days(i))
    fn = ['Daily Files', cyr, '.nc']
    Z = read_NetCDF(fn)
    t=Z.time
    x=Z.lat
    y=Z.long
    u=Z.u
    v=Z.v
    ii = find(x>=120 & x<=145)
    jj = find(y>=15 & y<=30)
    x=x(ii)
    y=y(jj)
    u = u(:, jj, ii)
    v = v(:, jj, ii)

ALSO

    ncfiles = dir('*.nc');
    Numfiles = length(ncfiles);
    all_U = cell(Numfiles, 2);
    all_V = cell(Numfiles, 1);
    
    %create array of all the u
    for i = 1:Numfiles
        all_U{i} = read_NetCDF(ncfiles(i).name, '-var', 'u');    
    end

Answer 1

[Vatsal]

Hi @Eleanor Alvarez,

I understand that you have daily files for three years, each containing five variables: "Lat", "Long", "Time", "U", and "V". The goal is to convert these daily files into weekly files. The "Lat" and "Long" variables will remain the same, while the "Time" variable will represent the week number (1 for the first week, 2 for the second week, and so on). Additionally, the "U" and "V" values will be updated with the average value for each week.

I am also attaching the code below to convert the daily files into weekly files: -

startDate = datetime('2008-05-01'); 
endDate = datetime('2011-04-31'); 
avgU = zeros(1440, 719); 
avgV = zeros(1440, 719); 

numFiles = 7; % Number of files to average per week 

numWeeks = weeks(endDate - startDate);

for i = 1:numWeeks 

    weekStartDate = startDate + (i-1)*7; 

    weekEndDate = weekStartDate + days(numFiles-1);

    sumU = zeros(1440, 719); 
    sumV = zeros(1440, 719); 
    filename = ['path_to_your_data/daily_file_', datestr(weekStartDate + (j-1), 'yyyymmdd'), '.nc']; 

    data = load(filename); 

    lat = ncread(filename, 'Lat'); 

    long = ncread(filename, 'Long'); 

    for j = 1:numFiles 

        % Load the daily file for the given date 

        filename = ['path_to_your_data/daily_file_', datestr(weekStartDate + (j-1), 'yyyymmdd'), '.nc']; 

        data = load(filename); 
        sumU = sumU + ncread(filename, 'U'); 

        sumV = sumV + ncread(filename, 'V'); 

    end 
    % Calculate the average U and V velocities for the week 

    avgU = sumU / numFiles; 
    avgV = sumV / numFiles; 

    % Save the weekly average data to a NetCDF file    

    outputFilename = ['path_to_save_weekly_files/weekly_file_', num2str(i), '.nc']; 

    nccreate(outputFilename, 'Lat', 'Dimensions', {'Lat', 719}); 

    nccreate(outputFilename, 'Lon', 'Dimensions', {'Long', 1440}); 

    nccreate(outputFilename, 'AvgU', 'Dimensions', {'Long', 1440, 'Lat', 719}); 

    nccreate(outputFilename, 'AvgV', 'Dimensions', {'Long', 1440, 'Lat', 719}); 

    nccreate(outputFilename, 'Time', 'Dimensions', {'Time', 1}); 

    ncwrite(outputFilename, 'Lat', lat); 

    ncwrite(outputFilename, 'Long', long); 

    ncwrite(outputFilename, 'AvgU', avgU); 

    ncwrite(outputFilename, 'AvgV', avgV); 

    ncwrite(outputFilename, 'Time', i); 

End 

You can also refer to the MATLAB documentation for "ncread" to obtain more information on its usage and syntax. The link is provided below: -

https://in.mathworks.com/help/matlab/ref/ncread.html

I hope this helps.

[Andebo Waza]

Dear all,

I need your help. I have wind data in 2D (Height, Time). I want to plot a 60-minute average data, and I want to do that using pcolor plot. I managed to generate the plot with the original data (30 min), but failed to do it on a 60-min average. I am relatively new to netcdf data. Any help is greatly appreciated!

Here is my code for original data:

clear all
close all
clc
ncid = netcdf.open('wl_geomar_CSM_l2_v04_20240916.nc','NC_NOWRITE');
filename = ncinfo('wl_geomar_CSM_l2_v04_20240916.nc');
disp(filename);
ncdisp('wl_geomar_CSM_l2_v04_20240916.nc');
longitude = ncread('wl_geomar_CSM_l2_v04_20240916.nc','lon');
latitude = ncread('wl_geomar_CSM_l2_v04_20240916.nc','lat');
time = ncread('wl_geomar_CSM_l2_v04_20240916.nc','time');
epoch = datetime(1970, 01, 01)
T = epoch + seconds(time)
u = ncread('wl_geomar_CSM_l2_v04_20240916.nc','u');
v = ncread('wl_geomar_CSM_l2_v04_20240916.nc','v');
wspeed = ncread('wl_geomar_CSM_l2_v04_20240916.nc','wspeed');
height = ncread('wl_geomar_CSM_l2_v04_20240916.nc','height');
u_mean=sqrt(u.^2+v.^2+wspeed.^2);
mymap=pcolor(T,height,u_mean);
shading interp
h= colorbar
caxis([0 20])
drawnow
ylabel(h,'Mean wind speed (m/s)');
grid on
xlabel('Time (UTC)')
ylabel('Height (m)')
xtk=xticks;

I would be happy to provide you the data if needed.

[Walter Roberson]

@Andebo Waza

I already answered at https://www.mathworks.com/matlabcentral/answers/1606970-how-to-plot-pcolor-in-matlab#comment_3319622