Granular computing is emerging as a new theory of computation based on representing and processing basic chunks of data, namely the information granules. Granular computing shares many concept and principles with the situation awareness paradigm, as both are focused on supporting the human capabilities of processing information, solving problems and making decisions. Accordingly, this work proposes a model for representing situations based on granular structures that sustains a computational approach in which granular computing mechanisms are used to support situation awareness. The model defines the concept of situation granularity as a measure to compare different situations, sustaining both comprehension and projection phases in the situation awareness assessment process. An illustrative example in the air traffic control domain demonstrates the validity of the proposed model.