Historic thermal calibration of Landsat 4 TM through an improved physics based approach

Show full item record

Title: Historic thermal calibration of Landsat 4 TM through an improved physics based approach
Author: Miller, Jonathan
Abstract: This work is intended to establish the calibration of the Thematic Mapper (TM) sensor, thermal band (band 6) aboard the Landsat 4 Spacecraft. Due to multiple organizations having operational control of the spacecraft and a lack of historical calibration data, a physics based approach will be used to calibrate the thermal data recorded over water during the operational lifetime of the spacecraft, which spanned 1983-1993. Using historical data from weather station observations, Radiosonde instruments, and moored weather buoys owned and operated by the National Data Buoy Center (NDBC), a ground truth comparison can be calculated and propagated through the atmosphere using a physics based model. The ground truth measurements are then compared to archived Landsat 4 data to determine how well the instrument is calibrated. This comparison over many data points allows construction of an overall calibration curve for the Landsat 4 data over the lifetime of the spacecraft. In addition to calibrating the Landsat 4 TM, this research will also include investigation into a possible transition period of negative to positive bias on the Landsat 5 TM, somewhere in the 1996-1999 timeframe. The same technique will be used in both calibration attempts. Results indicate that the Landsat 4 data was well calibrated in the timeframe before storage (1983-1984), however after the storage period (1987-1993) the sensor had a negative bias of -3.3 K. It is suggested to bias post 1987 Landsat 4 data by adding 0.4533 [W/m2 sr μm] to the sensor reaching radiance. Results from this correction radiometrically correct Landsat 4 data to ±0.48 K verified through an error analysis of the calibration procedure. The results from this study are consistent with the results from previous work performed on Landsat 5 TM in 2008. Results from the Landsat 5 study were inconclusive in finding a precise date for the bias shift in the 1996-1999 timeframe of Landsat 5 data. Twenty one additional data points were added to previous work completed in 2008, but no obvious shift was observed. This procedure for calibrating the Landsat 4 TM sensor tested the limits of the physics based calibration approach and proved that multiple buoys, multiple locations of the buoys (different bodies of water), unfavorable wind conditions, and Radiosonde data from larger distances away from the target can still produce valid results when calibrating a thermal sensor. Testing the robustness of the physics based calibration process opens the door to more available data, resulting in more extensive calibration curves for past work and future systems.
Record URI: http://hdl.handle.net/1850/12698
Date: 2010-06

Files in this item

Files Size Format View
JMillerThesis6-2010.pdf 6.720Mb PDF View/Open

The following license files are associated with this item:

This item appears in the following Collection(s)

Show full item record

Search RIT DML

Advanced Search