Comparison of Landsat Thematic Mapper and Airborne 
Thematic Mapper Simulator Data for Forest Damage 
Assessment 
C. Banninger 
Institute for Image Processing and Computer Graphics 
Joanneum Research, Graz, Austria 
Abstract 
Both Landsat Thematic Mapper (TM) and NS001 Thematic Mapper Simulator (TMS) 
data gave overall comparable results in discriminating stress conditions in a Norway spruce 
forest, with the higher spatial resolution NS001 TMS data more accurately delineating stressed 
canopy areas, but at the inclusion of numerous false anomalies associated with non stress- 
related canopy features. For canopies of high structural or compositional variability, the larger 
ground area covered by Landsat TM resolution cells mitigates the high spectral variability 
associated with small-scale irregularities in a canopy and thereby permits a more accurate 
assessment of the condition of a vegetation cover. 
Key Words: Landsat Thematic Mapper, NS001 Thematic Mapper Simulator, Norway 
Spruce Forest, Stress. 
Introduction 
Since its launch in 1982, Landsat Thematic Mapper (TM) data have been routinely 
employed in monitoring the state of health of forested regions world-wide. The sensor’s six 
reflectance bands covering the visible to shortwave infrard spectral region have proven partic 
ularly useful in detecting damage in forests of broad areal extent, such as the boreal forests 
of Canada and Scandinavia, but the 30 m spatial resolution of TM data has constrained their 
effectiveness and utilisation in the commonly mixed or mosaicked forests of central Europe, 
with their often broken or patchy canopy cover. Airborne sensor systems, such as the Daedalus 
Airborne Thematic Mapper (ATM) or NASA’s NS001 Thematic Mapper Simulator (TMS), 
provide equivalent Landsat TM spectral band coverage, but at a spatial resolution (generally 
5-15 m) that is more commensurate to the type, distribution, and areal extent of forests in 
central Europe. 
The ability of Landsat Thematic Mapper and NS001 Thematic Mapper Simulator data to 
discriminate between damaged and healthy trees in a stressed Norway spruce forest situated 
in southeastern Austria was evaluated with respect to TM and TMS data acquired during the 
mid-summer growing season. Single and transformed spectral data were co-registered with 
ground-collected data used to define the degree and extint of forest damagee at the study site 
and relationships derived between the different data sets to determine the effectiveness of each 
sensor system for damage assessment. 
857