In this paper we introduce an assessment and prediction technique for radar spectrum access in a dynamic electromagnetic environment. The proposed technique expands upon the existing spectrum sensing, multi-objective optimization (SSMO) framework for the joint optimization of the radar's signal to interference plus noise ratio (SINR) and range resolution. The proposed framework gathers training information in one spatial sector while the radar operates in another sector. The training information is used to form statistical estimates of the SINR and radio-frequency (RF) emitter activity. The predictive SSMO (pSSMO) technique then uses the training information during radar operation to avoid collisions with other RF emitters. Synthetic and measured Global System for Mobile (GSM) communication waveform data are processed by the proposed technique and the results indicate similar performance between the simulated and measured dataset, thereby validating the results.