GSA Annual Meeting, November 5-8, 2001

Paper No. 0
Presentation Time: 8:00 AM-12:00 PM

GEOLOGICAL CONTROLS ON VEGETATION REFLECTANCE: AN EXAMPLE FROM MIDDLE MAGDALENA BASIN, COLOMBIA


BAQUERO, Carlos Mauricio1, VINCENT, Robert K.1 and LEVINE, Norman S.2, (1)Dept. of Geology, Bowling Green State University, Bowling Green, OH 43403-0218, (2)Dept. of Geology, Bowling Green State Univ, Bowling Green, OH 43403-0218, baquero@bgnet.bgsu.edu

Geologic mapping in mountainous vegetated terranes is difficult under the best conditions. Techniques that aid in remotely mapping the geology in these regions are very important to natural resource development in many lesser-developed nations. This study investigates the use of remote sensing data for improved mapping of these difficult regions. An unsupervised classification of a Landsat 7 image showed strong correlation to the geology of approx. 5000 km2 of a rugged and densely vegetated area in the Middle Magdalena Basin, suggesting a link between geobotanical signals and the underlying geology. Removal of topographic and atmospheric effects is a necessary prerequisite before exclusively geology-vegetation relationships can be determined to occur. In order to isolate and enhance the geologic signals within the dense vegetation, band-ratioing and terrain corrections were employed. The analysis was aided by the use of digital topography and other geospatial data sets to enhance the classification of the region's geology. A field-based geological map was used to divide the image into homogeneous lithological polygons, allowing us to multispectrally assess the correlation between vegetation and lithology. GIS-based analysis allowed us to recognize analogies and dissimilarities between lithologies. Land-use polygons (tropical forest, rangeland and crops) were employed to constrain the analysis. If relationships between the Geology and vegetation prove to be consistent in the Middle Magdalena Basin, similar multispectral patterns may be useful for discrimination among broad lithological groups in other tropical basins. This study shows that the combination of remote sensing and GIS can be an extremely useful tool for the production of preliminary geologic maps from multispectral remote sensing data in vegetated areas.