A radar station (22) utilizes fuzzy logic processing to detect atmospheric microburst events. The detection process accesses radar velocity measurements to calculate wind shear and produce a wind shear likelihood image (500). Reflectivity measurements are used to create a reflectivity likelihood image as an indicator of storm activity (P52). A clutter image of stationary objects close to the radar is used to create a clutter likelihood image as an indicator of false alarms (P54). A weighting factor is applied to each set of wind shear, reflectivity, and clutter likelihood data, and the values are combined to produce a combined likelihood image (P56). The combined likelihood image is subjected to averaging over a fixed area to eliminate spurious readings, and positions having a sufficient size in combination with a likelihood value exceeding a minimum threshold amount are selected as microburst events (P58-P68).