Subset and nanmean of a three dimensional array to a have a 2d array 1xndim

Hi, I am very new to matlab and I need your help.

I have a 3d array 609x881x348 being lonxlatxtime of sst data.

I need to subset to a specific location, calculate mean (nanmean) and reshape this array in order to have a final 1x348 array, so having the nanmean of a subset of data in that specif area of interest for 348 time period.

So far I did this:

s1pos = (lon>=12 & lon<=16 & lat>=76 & lat<=77);  %609x881 logical of zeros and ones 
pt_lev(:,:,1) = nanmean(pt_lev8(s1pos)); 

But it doesn't give what am I looking for...

Hope the question is clear.

Thank you very much!!!

[SALAH ALRABEEI]

These coordinates looks like in the Barents Sea!!

[SALAH ALRABEEI]

 Your index work when the matrix is 2D; 
    % Try this example
clear
clc;
x=1:10; y=20:10:50;
[X,Y]=meshgrid(x,y); % if your lon and lat are already 2D ignore this step.
data=round(100*rand(length(x),length(y),3));
[dx,dy]=find(X<5 & X>2 & Y<50 & Y> 20);
dx = dx(1:2); dy([2,4]) = []; 
avg = squeeze(nanmean(nanmean(data(dx,dy,:),2)))

[CARLOTTA DENTICO]

Thank you!!

[Stephen23]

@CARLOTTA DENTICO: please accept the answer that helped you most.

http://web.archive.org/web/20210618082332/https://www.mathworks.com/matlabcentral/answers/858975-solved-subset-and-nanmean-of-a-three-dimensional-array-to-a-have-a-2d-array-1xndim

Answer 1

[Matt J]

pt_lev8=reshape(pt_lev8,[],348);

final = mean(pt_lev8(s1pos,:),1,'omitnan')

[CARLOTTA DENTICO]

Thank you very much, it works perfectly!!