ith 
| in 
ion 
ted 
cal 
Te 
its 
acy 
sult 
nal 
ion 
von 
nen 
ge 
der 
gen 
des 
satz 
len. 
ist 
Ben 
gen 
ung 
ner 
cial 
tive 
and 
the 
that 
nge 
Roeland de Kok 
1. Due to the rising population, the demand of timber in Tierra del Fuego has noticeably increased in recent years. 
Being a valuable hardwood, Nothofagus Pumilio could replace imports of Chilean Pinus radiata. 
2. Lenga Patagonia, of the American timber company Trillium that owns 57.000 hectares of argentine Nothofagus 
forest plans the construction of a sawmill and fiberboard plant. 
3. The owners of argentine sawmills want to expand their concession areas in order to be able to stay in competition. 
4. Duetoa cut in tax reductions by the argentine government, many foreign companies that were built in Ushuaia, the 
capital of Tierra del Fuego are closed. A sawmill could help reducing the increasing unemployment. 
In spite of heavy resistance by national as well as international environmental organizations, it will not be possible to 
stop an increased exploitation of Argentine Nothofagus forests. 
In order to assure the sustainability of the measures taken by timber companies, efficient and economic monitoring 
concepts are needed. 
As a consequence the goal of this work done as a diploma thesis in 1999 was to make statements about the possibilities 
of change detection in the Nothofagus forests and to create monitoring concepts using the given data material. 
2 DESCRIPTION OF THE STUDY AREA 
As a test area for the examinations, an area of about 50.000 hectares was chosen. It is situated in the central part of the 
Argentine part of Tierra del Fuego between the 52^ 30' S and 55^ 30' S of the South American continent. The province 
capital Ushuaia claims to be the southernmost city in the world. 
The area was chosen because specific studies were undertaken in this area during the PhD. thesis of Francisco A. 
Carabelli “Development of role models for a sustainable land use in Tierra del Fuego, Argentina”. 
All common forms of land use are found in this area as well as all of the important geomorphologic forms, soil types 
and vegetation forms. 
The vegetation forms found are the two forest types Nothofagus Antarctica and Nothofagus Pumilio that are found 
mainly in the south of Tierra del Fuego and the grassland, situated mainly in the northern region where it forms the 
southernmost parts of the vast grasslands of the Patagonian pampas (P. FREDERIKSEN 1988). 
Full coverage of satellite images as well as aerial photos was given. After two month of intensive ground-truthing work, 
enough information on the land cover and background information were gathered to start working on the image data. 
3 DATA MATERIAL 
The Material at hand was: 
e SPOT multispectral scenes from 1995, covering the whole argentine part of Tierra del Fuego. 
e Aerial Photos from 1960, covering most of the argentine part of Tierra del Fuego. 
e A vegetation map, digitized from the aerial photos, covering about 50.000 ha of forest 
e  GIS data covering infrastructure, waterbodies, ownership and provincial as well as state boundaries 
4 REMOTE SENSING DATA ANALYSIS 
One aspect of the work was to study the separability of the two different Nothofagus forest classes and the non-forest 
classes. First statements had to be made concerning the separability of the spectral signatures of the three classes. 
Therefore the SPOT image was classified using standard as well as object oriented classification procedures. 
For the change detection of the vegetation cover, datasets of the aerial photos combined with the three SPOT channels 
were produced and classified by means of object-based classification. 
Approaches commonly used for photo interpretation are pixel-based classification and/or extraction of information by a 
human analyst. Those two approaches to image interpretation have their own roles and often these are complementary. 
Whereas human interaction is good for spatial assessment but poor for quantitative accuracy, computer aided 
quantitative analysis has poor spatial ability and high quantitative accuracy (RICHARDS & JIA 1998). As works by 
SCHNEIDER, LICHTENWALD & AMMER (1996) show, those are the critical points when assessing change 
detection problems by a combined aerial photograph / satellite data approach. 
Using an object-based classification increases the computers spatial ability consequently making accurate computer 
aided photo interpretation possible. 
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B3. Amsterdam 2000. 215