In this paper we investigate a metacognitive radar (MCR) model that comprehensively combines disparate cognitive radar (CR) strategies for advanced performance in congested electromagnetic environments (EME). This model changes CR strategies as the spectral environment and target evolve for efficient radar dynamic spectrum access (DSA). The model first implements spectrum sensing followed by spectrum classification to identify known EME scenarios. These spectral scenarios assess the congestion and complexity of time-frequency data collected by the passive sensing process. This evaluation prioritizes possible CR strategies that are effective for the given spectral conditions. The MCR model then evaluates different CR strategies via learning and selects the technique that provides the best radar performance.