You can try the following, instead of "finding" the locations, which will be the indices, just use the locations like this:
sst = rand(10,20,30)*20;
ng=(sst>=15);
valuesAbove15 =sst(ng);
that would find all the values that are above 15. Notice that I simulated sst with random values of size [10 20 30].
if you then use the find:
ng2=find(sst>=15);
[row, col, page] = ind2sub(sz,ng2);
That would generate vectors of the locations of the same size as valuesAbove15, so you can now combine like this
mean(valuesAbove15(page==30))
and that would calculate the average values of the values above 15, that are in page==30. Just change to your values of interest and that would solve your problem hopefully.
