Errors included in Global Navigation Satellite System (GNSS) measurements degrade the performance of user position estimation but can be mitigated by spatial correlation properties. Augmentation systems providing correction data can be broadly categorized into State Space Representation (SSR) and Observation Space Representation (OSR) methods. The satellite-based cm-level augmentation service based on the SSR broadcasts correction data via satellite signals, unlike the traditional Real-Time Kinematic (RTK) and Network RTK methods, which use OSR. To provide a large amount of correction data via the limited bandwidth of the satellite communication, efficient message structure design considering service area, correction generation, and broadcast intervals is necessary. For systematic message design, it is necessary to analyze the influence of error components included in GNSS measurements. In this study, errors in satellite orbits, satellite clocks for GPS, Galileo, BeiDou, and QZSS satellite constellations ionospheric and tropospheric delays over one year were analyzed, and their spatial decorrelations and linear modeling characteristics were examined.