As dpb mentioned, if you have these minima calculated by matlab, it is easier to use the computed/saved information in that case. However, another solution is to use the find method, but to match the sizes as needed. For instance, I can make a random set of 3D data (A), find the minima (B), and then find the indices of each of the values. These indices are not recalled as a row and column, but rather as a the linear index of the original three-dimensional array.
for i = 1:10
A(:,:,i) = rand(3,3);
[B, Idx] = min(min(A));
C = find(A == repmat(B,[3,3,1]));
I am not exactly sure if that will work for you, or if you need the exact lat and long separately. If so, then this method probably would not work for you. However, it is an alternative that you might find useful if you really need to search for this at another time compared to calculation.