Expert systems have gained much recent attention as a means for capturing the performance of human experts in specialized fields of knowledge. Areas in which expert systems have been successfully developed include such varied applications as mass spectrogram interpretation, disease diagnosis, geological data analysis and computer configuration (Hayes-Roth et al, 1983). The assumption behind these applications is that a body of specialized knowledge is possessed by the human expert. Expert systems attempt to capture this knowledge in an explicit form, each as a set of heuristic rules, and employ mechanisms to apply this knowledge to solve problems in the domain of expertise. Using this approach, expert systems have been able to successfully perform tasks which previously could only be carried out by human specialists. Moreover, expert systems have in some cases been able to attain levels of performance equaling that of humans (Buchanan and Shortliffe, 1984). This paper describes an expert system-based approach to the problem of recognizing microbursts from Doppler weather radar data. A prototype system based on this approach is currently being developed at Lincoln Laboratory for automated recognition of low-altitude wind shear hazards. This system, called WX1, employs artificial intelligence and computer vision techniques to emulate the symbolic reasoning and visual processing capabilities of a radar meteorologist.