Version 40, 
Worcester. 
ham, 1993 
Habitat with 
neering and 
1038. 
d Reference 
tware, Inc. 
Landsat TM 
| and Trends 
ty of Civil 
95, 
20, Artificial 
\ Minimum 
I Supervised 
eering and 
73. 
LE. Mackey, 
early Cattail 
date SPOT 
ineering and 
25. 
Assessment 
dels, The 
al Wetland 
ct Technical 
. Fish and 
2, Wetlands 
> Data for 
nical Report 
Photography 
the United 
Vol. 33, No. 
chell, 1992, 
m Approach 
pp. 60-64. 
T.W. Sipes, 
| Swamps of 
S. Fish and 
; Program, 
GIS ANALYSIS OF INTEGRATED LANDSAT-TM, TOPOGRAPHIC, GEOLOGIC AND GEOPHYSICAL 
DATA SETS OF THE BASEMENT AREA OF THE RED SEA HILLS, SUDAN 
FRANZ K. LIST & NORBERT OTT 
Freie Universität Berlin, Institute of Geology, Geophysics und Geoinformatics, Berlin, Germany 
Commission VII, Working Group 4 
KEY WORDS: Geology, Landsat, Integration, GIS, Classification 
ABSTRACT 
Improved visualization and classification of lithologic units based on multispectral data is of major concern to geologists 
working with remotely sensed data. The results of spectral digital image analysis of data from current remote sensing 
systems are often disappointing due to the few and rather wide spectral bands available. Imaging spectrometry (e.g. 
AVIRIS, GERIS, HYDICE, AMSS) offers a solution for this problem, but hyperspectral data are still scarce and rarely 
available for the actual area of interest. Another solution is to use additional data sets together with the multispectral 
data, from existing geologic knowledge to geophysical and gechemical data. Classification of these disparate data sets 
is then performed in a GIS environment allowing to integrate different data types for improved lithologic discrimination. 
This approach is similar to the way in which a geologist, without help from a computer, would try to make sense out of 
the data available to him. Improved visualization of data is another topic with the human interpreter in mind, where data 
are arranged in a way analogous to our visual perception of the "real world". Examples from the Red Sea Hills, Sudan, 
serve to illustrate these points. 
KURZFASSUNG 
Eine Verbesserung der Visualisierung und Klassifizierung geologischer Einheiten auf der Basis multispektraler Daten ist 
für Geologen, die mit Fernerkundsdaten arbeiten, von grof$er Bedeutung. Die Ergebnisse spektraler digitaler Bildanalyse 
von Datensátzen üblicher Fernerkundungssysteme sind aufgrund der wenigen und sehr breitbandigen verfügbaren 
Spektralkanále oft enttàuschend. Abhilfe bieten Abbildende Spektrometer wie AVIRIS, GERIS, HYDICE oder AMSS: 
hyperspektrale Datensátze sind jedoch noch selten und für ein bestimmtes Arbeitsgebiet gewóhnlich nicht verfügbar. 
Einen anderen Lósungsweg stellt die Einbeziehung zusátzlicher Datensátze, von vorhandenem geologischem Wissen 
(z.B. geologische Karten) bis zu geophysikalischen und geochemischen Gelándedaten, dar. Eine Klassifizierung dieser 
unterschiedlichen Datensátze kann in einer GIS-Umgebung durchgeführt werden, in der verschiedenartige Datentypen 
miteinander in Beziehung gesetzt werden kónnen. Dieser Weg ist der klassischen Vorgehensweise eines Geologen 
verwandt, auch ohne Hilfe eines Rechners die verfügbaren Daten synoptisch zu interpretieren. Eine verbesserte 
Visualisierung geowissenschaftlicher Daten ist eine weitere Technik, durch die mit Hilfe eines Rechners Daten in einer 
Weise dargestellt werden, die der üblichen visuellen Erfassung unserer Welt durch den menschlichen Beobachter nahe 
kommt. Beispiele aus den Red Sea Hills, Sudan, belegen diese Aussagen. 
1 INTRODUCTION diorites and gabbros. The Precambrian faults became 
reactivated during the opening of the Red Sea in Pliocene 
11 GEOLOGIC SETTING times and represent block boundaries (SCHÓNFELD, 1977). 
The Red Sea Hills extend from Egypt into Sudan and Fig. 1 shows a georeferenced image mosaic of that area, 
Eritrea along the Red Sea, forming part of the Precam- produced from 5 Landsat TM frames. For the northeastern 
brian Arabian-Nubian Shield. A section of this highly com- part of the mosaic 2 Landsat MSS scenes had to be used 
Plex basement, situated at the Red Sea coast between since the coastal plain is cloud-covered on all available 
17 and 20° N, was being studied within a research pro- TM data. 
ject funded by the Deutsche Forschungsgemeinschaft 
(SFB 69). The study area consists of several terranes — The northern part of the mosaic is composed of low-grade 
which are divided by NE-SW-trending sutures. It is now metamorphic volcano-sedimentary and metasedimentary 
generally accepted that this situation points to an island- rocks, belonging to the Gebeit terrane. A prominent fea- 
arc assemblage with a number of ophiolite complexes ture in the mosaic is the NE-SW trending Nakasib suture 
(ALMOND ET AL., 1983; VAIL, 1988; SHACKLETON, 1994). that is cut and dislocated by the NNW-SSE trending Oko 
During the Pan-African episode (650-450 Ma) the rocks — shear zone. The Nakasib suture separates the Gebeit 
e metamorphosed under greenschist and amphibolite terrane from the Haya terran to the south. The metamor- 
acies conditions. This series known as "Greenschist As- phism of the metavolcanics and amphibolite/marble series 
Semblage" was intruded by abundant granitoid rocks, is higher than that of the Gebeit terrane. A second NE-SW 
425 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996