Sea ice detection from persistent single-channel shortwave infrared satellite data
The US Air Force has demonstrated an interest in deriving imagery products from classified military remote sensing platforms and making them available for civil and commercial operations. The US Air Force's Overhead Persistent Infrared (OPIR) is one such satellite constellation. A novel aspect...
| Published in: | Ecological Informatics |
|---|---|
| Main Authors: | , , , , , , |
| Format: | Text |
| Language: | unknown |
| Published: |
USMA Digital Commons
2019
|
| Subjects: | |
| Online Access: | https://digitalcommons.usmalibrary.org/usma_research_papers/150 https://doi.org/10.1016/j.ecoinf.2019.05.013 |
| Summary: | The US Air Force has demonstrated an interest in deriving imagery products from classified military remote sensing platforms and making them available for civil and commercial operations. The US Air Force's Overhead Persistent Infrared (OPIR) is one such satellite constellation. A novel aspect of OPIR imagery is its near-continuous capture of single channel shortwave infrared data over the Arctic. Although traditionally used for missile warning and strategic defense, the exceptionally high temporal resolution of the OPIR data stream makes it an attractive source for Arctic remote sensing, particularly as the Arctic has warmed at a rate nearly double that of lower latitudes. This work assesses the feasibility of using Geostationary Operational Environmental Satellite – 16 (GOES-16) data as a proxy for OPIR imagery in the Arctic. Specifically, we seek to determine whether a single channel shortwave infrared (SWIR) approach can be used to detect and chart Arctic sea ice. We used a series of 32-image daily sets (4 images per hour x 8 h) over four-day periods acquired by GOES-16 in late April 2016 (as well as mid-March, mid-May, and mid-June) to chart sea ice, clouds and water in Hudson Bay, Canada. To do this, we applied image enhancement techniques to raw data imagery and then employed a time-based classification algorithm to the enhanced data cube. Overall, our method successfully discriminated sea ice from water and clouds when all conditions were present with improved discrimination over current daily products for sea ice charting in the Northern Hemisphere. The simple methodology of the developed algorithm is critical to ensuring the temporal resolution of the sensor is capitalized. The rapid timeline for production of this type of data is essential to the relevancy to military operations as well as emergency response/preparedness operations in the Arctic as it becomes more accessible in coming years. Our results make a compelling argument for the application of Air Force Missile Warning data to assist in the mapping, tracking, and assessment of sea ice in the high Arctic. |
|---|