720 Routine Validation of the STAR Multi-Satellite Processing System Framework

Tuesday, 9 January 2018
Exhibit Hall 3 (ACC) (Austin, Texas)
William Straka, CIMSS/Univ. of Wisconsin, Madison, WI; and W. W. Wolf, S. Sampson, G. Quinn, R. Garcia, G. Martin, A. Li, M. Fan, J. Daniels, E. Schiffer, and A. DesMet

The GOES-R Algorithm Working Group (AWG) was initially tasked to develop and test candidate (Level 2) product algorithms in a scalable, pseudo-operational demonstration system, as well as develop validation and verification tools. In conjunction with NOAA/NESDIS/STAR, the Cooperative Institute for Meteorological Satellite Studies (CIMSS) at the University of Wisconsin has helped develop and validate algorithms in the GOES-R Algorithm Working Group (AWG) Product Processing System Framework. Routine validation of the various products as they are integrated, as well as performance characterization of the reference implementations, are important parts of this process. This processing system has evolved into the STAR Algorithm Processing Framework (SAPF), which is currently being used to not only validate output from the GOES-R Ground System but is also currently being run as the operational processing system for the VIIRS Level 2 products. To ensure that the algorithms are correctly integrated into the SAPF, tools were developed to compare datasets between several applicable product generation systems, allowing for quick comparisons that will identify issues or ensure that the algorithm was successfully integrated correctly. These same tools can and have been used to validate the implementation of the GOES-R AWG algorithms with the GOES-R Ground System as compared to the SAPF and various AWG teams. Once the algorithm is integrated in to the SAPF, a set of monitoring tools are used to visualize the products on a routine basis as well as provide statistics over time. These tools can alert the scientists when a given image produces unexpected results. The next step is routinely validating the products with truth datasets, such as the CALIOP lidar. Because the truth data is not in the same spatial and temporal step as the native satellite data, tools to collocate these data with satellite data and then validate the algorithms have been developed.

An overview and examples of these routine validation efforts for the enterprise algorithms produced by the SAPF from multiple instruments, including GOES-R imagery, VIIRS, Himawari-8 and other instruments as well, will be shown. In addition the tools used to do the validation will be discussed.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner