Verifying the Accuracy of Geostationary Weather Satellite Image Navigation and Registration

The next generation GOES-R geostationary weather satellites will provide imagery products with improved spatial and temporal resolutions and with more spectral bands than previous systems. Image Navigation and Registration (INR), which enables users to accurately pinpoint severe weather and stabilizes movie loops, will also improve. As INR performance improves, so must the technology for measuring INR performance. We describe our Product Monitoring (PM) system being deployed with the GOES-R ground system. It automatically measures INR performance using landmarks that are positioned with respect to a digital map created from the Shuttle Radar Topographic Mission (SRTM). Performance testing with Meteosat Second Generation (MSG) proxy data is part of the verification of the PM system, which is the main focus of this paper. A legacy system ironically called the Replacement Product Monitor (RPM) is in operational use on the GOES-NOP program. It is generally assumed that this system is capable of measuring the absolute position of landmark features relative to their mapped locations with an accuracy of about 0.5 pixels. This is plausible given that observed INR navigation error is about 1 pixel at the finest GOES-NOP resolution. However, a few landmark sites are observed to have biases possibly related to mapping error in the legacy digital map (not SRTM). Because the GOES-R system has finer spatial resolution than the GOES-NOP system and more stringent INR requirements, errors at the GOES-NOP pixel level are quite important. Our verification work with the GOES-R PM seeks to systematically characterize the measurement errors in a controlled test environment to demonstrate its suitability for a GOES-R mission with finer spatial resolution and more stringent INR requirements in comparison with GOES-NOP.