Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4972800 | ISPRS Journal of Photogrammetry and Remote Sensing | 2017 | 17 Pages |
Abstract
Two independent VIIRS-based Sea Ice Concentration (SIC) products are validated against SIC as estimated from Very High Spatial Resolution Imagery for several VIIRS overpasses. The 375Â m resolution VIIRS SIC from the Interface Data Processing Segment (IDPS) SIC algorithm is compared against estimates made from 2Â m DigitalGlobe (DG) WorldView-2 imagery and also against estimates created from 10Â cm Digital Mapping System (DMS) camera imagery. The 750Â m VIIRS SIC from the Enterprise SIC algorithm is compared against DG imagery. The IDPS vs. DG comparisons reveal that, due to algorithm issues, many of the IDPS SIC retrievals were falsely assigned ice-free values when the pixel was clearly over ice. These false values increased the validation bias and RMS statistics. The IDPS vs. DMS comparisons were largely over ice-covered regions and did not demonstrate the false retrieval issue. The validation results show that products from both the IDPS and Enterprise algorithms were within or very close to the 10% accuracy (bias) specifications in both the non-melting and melting conditions, but only products from the Enterprise algorithm met the 25% specifications for the uncertainty (RMS).
Related Topics
Physical Sciences and Engineering
Computer Science
Information Systems
Authors
Daniel Baldwin, Mark Tschudi, Fabio Pacifici, Yinghui Liu,