How can I average data from June,July, and August into one matrix

I have monthly data files that are 768x1152 (latitudexlongitude). I am reading in 3 monthly files from 1980 through 2005. I want to average the 3 monthly files for each year and put the average into a matrix that is latitude by longitude by time. With time being the 26 years.

Here is my code currently:

lon = ncread('/glade/p/cesm/amwg/runs/cam5_1_amip_run2/cam5_1_amip_run2.cam2.h0.1980-01.nc','lon');
lat = ncread('/glade/p/cesm/amwg/runs/cam5_1_amip_run2/cam5_1_amip_run2.cam2.h0.1980-01.nc','lat');
% Read in lat lon data

klat = find((lat>=0)&(lat<=40));
klong = find((lon>=280));
klat = klat(1:3:end);
klong = klong(1:4:end);

% Find lat/lon for only the North Atlantic and skip points

pressure = NaN(length(klat),length(klong),26);
% Create an array of NaN to fill in loop

dir = '/glade/p/cesm/amwg/runs/cam5_1_amip_run2/cam5_1_amip_run2.cam2.h0';

for y =1980:2005
   for m =6:8
     year=num2str(y);
     month = num2str(m);
     infile=[dir '.' year '-0' month '.nc'];
     infile1 = [dir '.' year '-0' month '.nc'];
     infile2 = [dir '.' year '-0' month '.nc'];
     June = ncread(infile,'PSL');
     June = June'; % needs to be transposed so it is lat by long
     July = ncread(infile1, 'PSL');
     July = July';
     August = ncread(infile2, 'PSL');
     August = August';
     JJA = [June July August];
     mJJA = mean(JJA);
     index=y-1979;
     pressure(:,:,index)=mJJA(klat,klong);
   end
end
% Create loop to read in 26 years of average data from June, July, and August

press= reshape(press,3648,26);
% Reshape data so it is 2 dimensional

So the problem I have is when I go to save the averages into the array pressure I get an error because the averages are (1 by 3456) whereas klat by klong is 3648 as shown in the reshape.

[Ced]

Hi

Your average does not really make sense to me. What you want is to have a pressure array for each latitude and longitude point, averaged over the three data sets of each month, correct?

Because what you are doing is taking the files of each month, and then averaging over the longitude. (That's why you have a vector, and not a matrix as you probably want)

Then, for the dimension error: It seems as if the data you are reading in actually has 1152 longitude points per month, and not 1216 as you are expecting. At least that's what I can gather from your comment.

In short, a few comments:

1) for the dimension problem: check again for which latitude/longitude points your data is, and form klat, klong accordingly

2) Maybe I'm missing something, but you seem to be reading the same file three times (infile == infile1 == infile2) ??

3) If you want to average each month, I would do something like

nlat = length(klat);
nlong = length(klong);

pressure = zeros(klat,klong,26); % nicer than NaN imho
June = zeros(klat,klong,3); 
July = zeros(klat,klong,3);
August = zeros(klat,klong,3);

for i = 1:26 % for each year
    for j = 1:3 % for each data set
        infile = name_for_data_set_j_of_June;
        June(:,:,i) = ncread(infile);

          % ... repeat for july and august
    end

    % now average over the three data sets and save
    pressure(:,:,i) = mean([June July August],3);
end

So yeah... There are quite a few points which don't really make sense to me in how your read and save your data, but maybe that will help you move forward.

Answer 1

[Chad Greene]

If I understand your question correctly, you should be able to use downsample_ts.