A “microarray” as defined in this paper is modeled on a sub-geometry of the NORESS array, and comprises a 3-component center seismometer surrounded by 3 closely spaced vertical-component sensors deployed over a typical aperture of 0.3 km. Analysis of five days of continuous data has shown that such a system combines the benefits of array and 3-component processing in providing reliable automatic detection, phase identification and azimuth estimation of weak seismic events at local and regional distances. Using vertical components only, broadband f-k array analysis enables correct phase identification (P-type or S-type phase) in 95% of the cases, and gives S-wave azimuths with a root-mean-square (RMS) deviation of 13.9 degrees from the estimates of the full NORESS array. It is particularly significant that the small array eliminates the need for introducing particle motion models, which creates ambiguities in 3-component analysis of secondary phases when interfering SH and SV phases occur. P-phase azimuths are estimated using integrated array and 3-component f-k analysis, and have an RMS deviation relative to NORESS of only 9.6 degrees. Compared to the full NORESS array, the P-wave detection capability is reduced, especially for epicentral distances greater than 500 km. The S-phase detection capability is enhanced by incoherent beamforming of the horizontal channels, and approaches that of NORESS at all distances. A considerable reduction in the detector false alarm rate is achieved by imposing constraints on the estimates of apparent velocity obtained from the f-k analysis before accepting a detected phase.