The Landsat-8 Data Users Handbook provides information on some of the common known OLI and TIRS instrument image artifacts.

In addition, USGS maintains a list of L8 OLI and TIRS calibration notices that describe anomalies affecting the sensors and the L8 products. This list provides information on new or temporary artifacts, how they are addressed and corrected, and provides up-to-date information and details regarding known issues and current artifacts.
 

TIRS Anomaly

Image quality issues have been detected on the TIRS ESA products which systematically contaminate the Band 10 (Figure 1) and Band 11 (Figure 2) TIRS data: image striping (at the detector level) and image banding (at the camera level).

Figure 1 : LC82000342017058MTI00, Band 10

Figure 2 : LC82000342017058MTI00, Band 11

Figure 3: LC82000342017058MTI00 – Band 4, Band 10, Band 11 composite – Left : full image, Right : ROI.

The image striping is due to a detector-to-detector mis-calibration (relative calibration). The recovery plan involves the reprocessing of the data with a new set of band adjustment coefficients that were proposed by the USGS in March 2014 and new Response Linearization Look Up Tables (RLUT) parameters that were released in March 2015

For information, all ESA products acquired after April 11 2017, are processed with a more up-to-date version of the RLUT (‘L8RLUT20150303_20431231v11.h5') which corrects mis-calibration, as shown in Figure 4 and Figure 5 below. The RLUT information can be found in the product metadata file at the ‘RLUT_FILE_NAME' parameter.

Figure 4 : LC82000342017122MTI00 – Band 4, Band 10, Band 11 composite – Full image.

Figure 5 : LC82000342017122MTI00 – Band 4, Band 10, Band 11 composite – ROI.

The image banding is due to stray light and the abrupt change in absolute calibration. An operational stray light correction algorithm has been developed by the USGS whereby "results show that the magnitude of the banding artifact is reduced by half on average over the current (uncorrected) product and that the absolute radiometric error is reduced to approximately 0.5% in both spectral bands on average (well below the 2% requirement)." This effect is not systematic and a software update is required to improve the ESA products.

The above issues do not affect the OLI data; however discussions are ongoing with the view to reprocess the ESA Landsat-8 data with the latest software and supporting data in order to recover the affected products.