Sum the daily data based on month and year

I have a three dimensional matrix of gridded data (16x18x7669) where 16 is the latitudes, 18 the longitudes and 7669 is the daily data from 1/1/1998 to 31/12/2018. I would like to get the monthly and yearly totals for each grid. I am a newbie in Matlab and I will appreciate suggestions on how to go about it. Thank you in advance

Answer 1

[Stephane Dauvillier]

Lets assume your 3D variable is called data

Frist get the group id : which width of data correspond to the same month of the same year:

time=datetime(1998,1,1):datetime(2018,12,31);
grp = findgroups(year(time),month(time));

Then cget the unique list of group

uniqueGr = unique(grp);

Initialize the result and loop on unique groupe element

result = zeros(size(data,1),size(data,2),numel(uniqueGr)) ;
for iGroup = 1:numel(uniqueGr)

Compute the sum on the 3rd dimension but only for the data which corresponding date to the current group

    result(:,:,iGroup) = sum(data(:,:,grp==uniqueGr(iGroup)),3);
end

Note there is the function splitapply that makes this easier if you have 2D array

[Ryn]

Hi Stephane,

This works but Iam getting unrealistic values and I cannot figure out what I am doing wrong. This is what I have tried for the yearly totals:

clear;close all;clc;
load CHIRPS2
%%yearlytotals
yGroup = findgroups(year(dateCHIRPS));
 uniqueGrYr=unique(yGroup);
 
%Initialize the result and loop on unique groupe element
pCHIRPS_Y=zeros(size(precipCHIRPS2,1),size(precipCHIRPS2,2),numel(uniqueGrYr)); 
for iGroup = 1:numel(uniqueGrYr)
      
    pCHIRPS_Y(:,:, iGroup) = sum (precipCHIRPS2 (:,:, yGroup == uniqueGrYr (yGroup)), 3);
end

[Stephane Dauvillier]

What do you mean by "unrealistic values" ?

You ask to sum the whatever data you have for each year.

Let's just take a tiny example

years = [1990,1990,1991];

data(:,:,1) = [1,2;3,4;5,6] ;
data(:,:,2) = [7,8;9,10;11,12];
data(:,:,3) = [13,14;15,16;17,18];

Then you will have as a result

for year 1990

result(:,:,1) = data(:,:,1) + data(:,:,2) = [8,10;12,14;16,18]

and for year 1991

result(:,:,2) = data(:,:,3) = [13,14;15,16;17,18];

[Ryn]

Thank you Stephane

Answer 2

[ABHILASH SINGH]

You can try this;

https://in.mathworks.com/matlabcentral/fileexchange/74497-daily-data-to-monthly-average-sum-in-matlab?s_tid=LandingPageTabfx