An Enhanced Automated Process to Support Landsat Thermal Band Absolute Calibration

Abstract:

The need for accurate ground truth is important for any remote sensing platform in order to validate calibration methodologies. For thermal infrared imagery, large bodies of water are suitable as ground targets for this purpose. It has been demonstrated for both MODIS and Landsat sensors that temperature measurements from floating buoys maintained by NOAA along with atmospheric information provided by radiosonde profile data can be used to model top-of-atmosphere calibration truth. However, this process is highly sensitive to accurate atmospheric knowledge. This work improves upon the previous methodologies by substituting North American Regional Re-analysis (NARR) data for radiosonde profiles in the top-of-atmosphere modeling process. A streamlined procedure is developed to automatically ingest NOAA buoy temperature data along with the appropriate NARR atmosphere data in order to model the expected top-of-atmosphere radiance for the given sensor (Landsat 8/TIRS in this case).