Daily to monthly values in a loop

Hello,

I have 50 columns of monthly data for the years 1951-2000. However, the monthly data is in fact the daily average for that month and now I would like to convert it into monthly total by multiplying the daily average by the number of days in each month. So I want to multiple every January in the time series by 31, February by 28, March by 31 and so on.

I would like to write a loop that processes all months and all columns but am not quite sure how I can tell matlab to multiple each month with the correct number of days. Any help would be greatly appreciated :)

Answer 1

[Andrei Bobrov]

yourdata - your array (12 x 50)

out = yourdata.*bsxfun(@(x,y)eomday(x,y),1951:2000,(1:12)')

corrected

yourdata - your array (600 x 50)

dsmhyr = bsxfun(@(x,y)eomday(x,y),1951:2000,(1:12)');
out = bsxfun(@times,dsmhyr(:),yourdata);

more variant

y=1951:2000;
N = numel(y)*12;
daysinmonth = 31*ones(N,1);
daysinmonth(bsxfun(@plus,[2:2:6,9:2:12]',0:12:N-1))=30;
n = 2:12:N;
daysinmonth(n(~rem(y,400)|(rem(y,100)&~rem(y,4))))=29;
out = bsxfun(@times,daysinmonth,yourdata);

[Anna]

Thanks so much for your answer!

I've still got a slight problem with the code though. my array is in fact (600 x 50) because I have monthly data for 50 years. How can I repeat the above command for all the 600 rows in each column?

[Andrei Bobrov]

in your array (600x50)

column - year?

row - ???

[Anna]

column: monthly runoff data for 50 different river basins (so each column represents a different basin)

row: time series of data

row 1: runoff in Jan 1951

row 2: runoff in Feb 1951

row 3: runoff in Mar 1951

...

...

row 600: runoff in Dec 2000

[Andrei Bobrov]

see corrected