Identifying zones in fluid domain with max velocity gradients

Hi

I am exporting CFD data from FLUENT in the ASCII format to be read by a MATLAB algorithm which generates reduced order approximations of complete flow fields at new boundary conditions. I am working with natural convection in a differentially heated cavity in 2D and have to find a correlation with velocity field to surrounding thermal boundary conditions.

I need to find a way for my algorithm to automatically locate a finite set of arbitrary control surfaces in the fluid domain that corresponds to the maximum velocity gradients for the x and y components of velocity. It may be acceptable for me to predefine the length of the surfaces dx and dy, which should make things easier. The offline data is stored in a matrix and I have the node numbers, x and y coordinates and velocity components of the CFD solutions at my disposal. The x and y components and corresponding surfaces are to be evaluated separately. Output: column vectors of node numbers for:

[1] a horizontal surface for y component subject to max(dvy/dx)

[2] a vertical surface for the x component subject to max(dvx/dy)

To date I have been doing this based upon guess work in the script with the find function, but results will improve drastically if I can get the script to generate the surfaces subject to the above constraints.