Landsat 5 anomalies

This page provides a brief description of the possible anomalies that can affect Landsat5 TM products, in-depth explanations can be viewed on the ESA or USGS web sites.

Swath misalignment

Swath misalignment due to CPF

The calibration parameter file (CPF) supplies the radiometric and geometric correction parameters required during Level 1 (L1) processing to create products of uniform consistency across the Landsat system.

The Image Assessment System (IAS), an element of the Landsat Ground Segment, periodically performs radiometric and geometric calibration and updates the CPF, which is labelled with applicability dates and sent to international ground stations via the Landsat Mission Operations Centre.

If the CPF file has not been updated by the Ground station your Landsat TM product may be affected by swath misalignment.

Swath misalignment (magnified)

How to verify if the CPF has been updated:

For FAST products, the CPF file name is described in the metadata file ("*.mtl"); line 55, field "CPF_FILE_NAME" where group is "PRODUCT METADATA".

For CEOS products, the same information is provided in the volume directory file of the calibration file as in the image below.

A VDF_CAL.001 file opened with a text editor, with the date range highlighted in red

If the acquisition date of the Landsat 5 TM product that you purchased is within the date range given, the CPF file has been updated and the scene is not affected by swath misalignment

Swath misalignment due to detector displacement

Detector displacement patterns concentrated on a specific image part are due to geometric calibration ("resampling") and the use of Nearest Neighbour option (it leads to small swath shifts see Figure A below).

Small swath shifts are allowed up to one (1) pixel - see this ESA report

In some cases, the use of another geometric calibration method (different to Nearest Neighbour) might solve these detector displacements (e.g. Cubic Convolution method, magnitude is up to 1 pixels)

Striping

Striping anomalies occur at scene "line" level and are due to detector-to-detector mis-calibration. This ESA report clarifies the main causes of this anomaly and assess its consequences on radiometric product quality. Also for this anomaly the use of another geometric calibration method (different to Nearest Neighbour) might solve these detector displacements (Cubic Convolution method is suggested).

Ephemeris and attitude

Product geometric accuracy is in part strongly dependent on ephemeris and attitude metadata reliability. This ESA report deals with anomalies due to inconsistent and shifted ephemeris and attitude records and their consequences on product location accuracy.

Missing image data

Anomalies due to missing image data such as missing pixels, missing swath - speckle .... Root causes differ strongly from one anomaly to the other one. In some cases real pixel values can be recovered. This ESA document provides the user with basic steps for implementing a repair process.

Banding

This ESA report addresses common radiometric artefacts mainly occurring at scene swath level on Landsat 5 TMs and classified as memory effect, scan line correlated shift and coherent noise.

Detector oversaturation

This anomaly, commonly known as the "comet-like" artifact, is caused by a very bright object on the ground that exceeds the dynamic range of the instrument. This ESA report proposes precise enquiries for a better understanding of the phenomenon.