If you are sure that Range-Azimuth is obtained via a 1-D FFT, then you can just do an inverse FFT, as long as you have the original complex Range-Azimuth response, not just the magnitude from the plot.
However in general this information is not available. So one possible way to get around is to extract the directions and then simulate the data. For example, you can use sensorsig in Phased Array System Toolbox to simulate the test signal.
HTH